THE INFLUENCE OF CLIENT VARIABLES ON PsYCHOTHERAPY Jou F. Carin ‘New York Hospital, Cornell Medical Center Kenner N. Levy ity University of New York INTRODUCTION Controversy about the effectiveness of psy- chotherapy has a long history. In 1952, British ‘experimental psychologist, Hans Eysenck, caused a furor when he proclaimed that the application of psychotherapy was no more beneficial than the absence of treatment. In is report, Eysenck (1952) summarized the results of 24 reports of psycho- analytic and eclectic psychotherapies with more than 7,000 neurotic clients treated in nacuralis- tic settings compared with two! éontrol groups. Eysenck found that the more intensive the ther~ apy, the worse the results. In fact, Eysenck’s data suggested that clients in psychoanalytic treatment hhad significantly worse cure rates than clients who received no treatment. thas been more than 40 years since Eysenck rocked the treatment community with his claims that psychotherapy did not work. Despite the use of seriously flawed research methodology and 2 polemic tone, Eysenck’s article was extremely important to the field and challenged clinical psy- chologists to pay more systematic attention to the results of their efforts and has spurred a great deal of empirical research. ‘Thanks in large part to researchers’ responses to Eysencks charge, we now know, generally speaking, that psychotherapy does indeed help people (Lambert, Shapiro, & Bergin, 1986; Smith, Glass, & Miller, 1980; Chapter 5 in this volume). Numerous studies and subsequent meta-analyses have demonstrated that any number of specific psychotherapeutic approaches, ether alone or, in some cases, in combination with pharmacologi- cal approaches, are more effective than credit alternative psychological interventions containing nonspecific factors serving as “psychological plac bos" Barlow, 1996). that psychotherapy works; nevertheless, psy therapy research is ata critical period. A conue of pressures both inside (e.g. evidence-sup treatment movement) and outside the pro tion, National Alliance for the Mentally 111) it incumbent upon clinical psychologists to becom research efforts on more precise questions, such a Given a clients diagnosis, which treatment is rec ‘ommended? What treatments have shown effi in empirical trials? Does the therapy prodi ments that show efficacy in clinical trials b demonstrated similar effectiveness in local teat ‘ment setting? : “The issue of client variables is an abstract unique in terms of range and severity of problems developmental history and achievements, inte personal skis, intellectual acumen, state of paid) and desire for change. Many characteristics of the client may potentially influence the therapevtt] venture. At the same time, the clients! behavior therapytwill be influenced by the characteris 194dible lace agree orted a make some sy sing icy duce nges dure reat have reat ats ems, er sin, fhe tie arin stics and behavior of the individual therapist, for the therapeutic process is basically an interpersonal phenomenon. With a deceptively simple wisdom, Jerome Frank (1973) pointed out long ago that psychotherapy is an encountér between a demor- alized client and a therapist whose goal is to ener- gize the other. These straightforward truths lead tus to the more refined questions: Which client and therapist characteristics interact most saliently and forcefully to produce symptom decline? Which of these interactions lead to improved social and ‘work adjustment? Comparative outcome studies of psycho- therapy are costly and time consuming, and for the most part have not yielded clear evidence of the superiority of specific psychotherapies for specific disorders. Recent psychotherapy research has focused on the client's “diagnosis” and the techniques of therapy while ignoring the idiosyn- cratic aspects of the client that are even more salient in predicting change and guiding treat- ‘ment decisions. However, large-scale studies com- paring different forms of treatment for different disorders have revealed few differences in out- come based on technique. For example, recent examinations of psychotherapy outcome and process in the Treatment of Depression Collabo- ative Research Program (TDCRP) suggested that outcome is better predicted by client charac- teristics than by the effects of particular kinds of interventions (Ablon & Jones, 1999; Blatt, Quin- hn, Pilkonis, & Shea, 1995; Zuroff et al., 2000). Reviewers (Bergin & Lambert, 1979; Frank, 1979) have suggested that the largest proportion of vari- ance in therapy outcome is accounted for by the personal characteristics and qualities of the client. As much as 40% of client improvement in psy- chotherapy can be attributed to client variables and extratherapeutic influences (Lambert, 1992). ‘These findings suggest that the study of client variables may have much to offer for our under- standing of psychotherapy’s effectiveness. Ident ication of premorbid clinical and personality characteristics predictive of outcome might help dinicians guide treatment choices and revise ‘Teatment methods based on the needs of dif nt types of clients. ‘This chapter highlights the client attributes ‘and characteristics that profoundly shape and influ- nee therapeutic process and outcome. We review 4 number of relevant conceptual and method- logical issues related to the influence of client ‘ariables on therapy selection, processes, and out- ‘Some. This chapter builds on the previous edi- ‘The Range of Client Characteristics « 195 tions of this chapter (Garfield, 1994; Garfield & Bergin, 1986). Garfield’ (1994) ast review empha- sized client variables in isolation, whereas we think the field is currently emphasizing client variables in interaction with both therapist and treatment variables. We emphasize client vari- ables as mediators and moderators of psychother- apy process and outcome. Throughout this review ‘we emphasize the interaction of client character- istics and the growing relationship with the ther- apist. This interaction is such that any research focused exclusively on client variables is (falsely, in our minds) assuming thatthe therapist reaction does not influence the client varidble in question As research in this area becomes more sophisti- cated, the interaction of client characteristics with therapist response will likely become the focus of clinical concem and research interest. ‘The previous chapter emphasized specific client variables of social class, personality, diag- nosis, age, sex, intelligence, and length of dis- turbance. In this chapter, we review more current constructs relating to the client such as inter- personal relatedness and preparation for change. Since the previous edition (Garfield, 1994), psy- chotherapy research data by client diagnosis has grown considerably. Although this orientation hhas its strengths and weaknesses, the accumula- tion of data organized and investigated by client diagnosis and related treatment is so prominent in the field that it necessitates some review in this chapter. Garfield (1994) mentioned the luence of socioeconomic variables and eth- nicity, but there has been a major accumulation of data on the psychotherapeutic influence on clients, with a diversity of ethnicity and socioe- conomic levels, and we emphasize findings in this area. ‘THE RANGE OF CLIENT CHARACTERISTICS ‘The number of client variables with potential for informing the process and outcome of psy- chotherapy is virtually limitless. Everything from ‘genome and brain chemistry to demographic va ables and environmental conditions to personality ‘traits, to problem area/diagnosis is arguably related to psychotherapy and its ingredients. Client char- acteristics can be external to the individual (e.g, social support) or intimate aspects of the individ ual (€g., intelligence). Client characteristics can be invariant (e.g, gender, ethnic membership), relatively stable (eg., SES, personality traits), or196 + Chapter 6 / The Influence of Client Variables on Psychotherapy quite variable (e.g., motivation for change). Client variables can be psychological in nature such as personality traits, or they can be part of the indi- viduals’ biological system (e.g., state of REM sleep characteristics). Over the years the type of client variables investigated has apparently shifted from stable-demographic variables to a broader range of variables, with increasing emphasis on the interaction of client variables with treatment variables as provided by the therapist. ‘The presence of an almost limitless number of client variables forces the reviewer (and clini cian) to select those variables thit have proven ‘most relevant to essential aspects|of the therapy enterprise. With the advantage of growing body of information on the key processes and outcome ‘of psychotherapy research, we hive elected to focus on the specific client variable that relate to the matching of client and psychotHerapy, process of psychotherapy, and therapy outcome. Not only are there different types and sources of client variables, but these variables function in different ways in relation to psy- chotherapy process and outcome. Client variables can be conceptualized as static predictors of response to treatment. Thus, the clients’ gender or ethnic membership can be examined as a pre~ dictor of treatment process or outcome. A client variable can be seen as a moderator or mediator of change (Holmbeck, 1998). A mdtlerator vari- able affects the relationship betwee the predic- tor variable and a dependent variable, and the value or level of the moderator varidble makes a differential impact on the dependent variable. In contrast, a mediator variable is a| mechanism through which the independent varjable affects the dependent variable. Thus, the independent variable influences the mediator, which, in turn, influences the outcome or dependent variable. Finally a client variable can be conceptualized as a prescriptive variable, that is, ong that pre- scribes a certain treatment as opposed|to compet- ing treatments. For example, in the evidence-based treatment movement, the client variable of diagnosis is seen as a prescription for certain psychotherapies. CHARACTERISTICS OF THOSE WHO SEEK THERAPY Those Who Seek Therapy ‘Who is the psychotherapy client? What are the characteristics of individuals who request or receive psychotherapy in contrast to those who do not? Our examination of client variables ang psychotherapy matching, process, and outcony should not be limited to the client variably describing only those who undergo psychothet apy. However, knowledge of those who obtaig psychotherapy does help define the limits ofthe current research information on client variable and psychotherapy. : In the general population, those who repon emotional distress (Veroff, Kulks, & Douvan, 1981; Ware, Manning, Duan, Wells, & New. house, 1984), exhibit ‘psychological symptong Boyd, 1986; Yokopenic, Clark, & Aneshense), 1983), and consider their mental health to br poor (Leaf et al., 1985) are most inclined to seet professional mental health care. Women are more likely than men to seek both informal support and professional help (Butler, Giordano, & Nerea, 1985; Horwitz, 1977; Kessler, Brown, & Broman, 1981). Age is also related to help-secking behav. jot. The elderly are more reluctant than younger individuals to seek help from mental health pro: fessionals, and they rely more readily on generd medical practitioners and the clergy (Leaf, Bruce, Tischler, & Holzer, 1987; Waxman, Carer, & Klein, 1984). Those elderly who sought assis tance, as compared to those who did not, had poor psychological well-being, more physicd health problems, a higher level of stressful events, and greater deficits in social support (Phillips & Murrell, 1994) Stress is related to seeking the services of mental health professionals, though in a some- what complicated manner. Not everyone who experiences stress seeks mental health services. ‘Those seeking assistance may experience the impact ofthe stressors more intensely (Goodman, Sewell, & Jampol, 1984) and are less likely 10 hhave strong social support from friends and rel tives (Birkel & Reappucci, 1983). Howard and colleagues (Howard etal, 1999 have summarized patterns of mental health ser ice utilization using data from the Epidemislogc Catchment Area (ECA) survey and the National Comorbidity Survey (NCS). Both studies indi- cated that about 30% of adults will experience # diagnosable mental condition in any given yea and the majority of these individuals (from 56 t@ (60%) will have more than one disorder. Whats striking is that in the ECA survey, more than 70% of those with a mental disorder received no ser” ices, and only 13% obtained treatment from # mental healgh professional. This would indicat that the vast majority of data we have on cliensi | andits relationship to psychotherapy is based on t | information from a very small percentage of the + | individuals who actually need intervention, farly Termination t | Barly termination or attrition from psychother- + | gpyisan issue that has important clinical implica- tors. From the clinician’ point of view, those + | indivduals who drop out from treatment prema- 1 | tuely are not taking advantage of an important 2] feourse in their lives. Ifthe early termination s | can be predicted, the initiation and course of 1. | therapy can potentially be modified in order to | motivate the client for concentrated work toward k | change and a reduction in premature dropout. : ‘Most studies have suggested that age is not J | important in psychotherapy retention (DuBrin & 4 | Zastowny, 1988; Gunderson et al., 1989; Sledge, 4 | Moras, Hardey, & Levine, 1990). In contrast, 2 | several other variables seem to be important. In a + | multisite study of panic disorder, the client vari- ables of lower household income and negative 1 | asides toward the treatment offered were inde- + | pendently associated with attrition (Grilo et aly 1998). Similarly, it was found that for clients ~ | suffering from obsessive-compulsive disorder, 1 | strong incongruent treatment expectations pre- 1 | diced attrition (Hansen, Hoogduin, Schaap, & de Haan, 1992). Organista, Munoz, and Gonzalez (1994) and Mirands, Azocar, Organista, Dwyer, and Arean f | (ander review) evaluated the benefits of a group ~ | cognitiveybehavioral treatment for depression in » | lients with low income and the majority of whom vere from Latino or African-American minority 2 | groups. The dropout rate was higher in this low- ‘income minority population (40 to 60%) than in » | the NIMH multisite depression study (Elkin, Shea tal, 1989) Importantly, Miranda and colleagues found that adding case management services sig- nificantly reduced the dropout rate. Significane improvement in depression was reported in both | Seles, baton average the clients remained inthe i pressed range even with treatment according t0 ‘heir self-report questionnaire information. onlin witha personality disorder diagnosis re i fon ob a high ik for premae out, whether in inpatient settings (Chiess, Drahorad, & Longo, 2000) or outpatient treat- ‘ent settings (Gunderson etal, 1989; Shea et al. nad Skodol, Buckley, & Charles, 1983). The dropout rates vary from 42% (Gunderson etal, '989) to 67% (Skodol, Buckley, & Charles, 1983). ven the dropout rate, the question becomes one Characteristics of Those Who Seek Iberapy * 271 of understanding the operative variables. Clarkin and colleagues (Smith, Koenigsberg, Yeomans, Clarkin, & Selzer, 1995; Yeomans etal, 1994) ana- fyzed factors associated with attrition from psy- chotherapy for clients diagnosed with borderline personality disorder. They found that younger clients and those with high initial hostility were more likely to withdraw early from treatment. Ina subset of clients, those who showed a predomi- nance of narcissistic themes in their responses on the Rorschach test at the beginning of treatment ‘were more likely to drop outof treatment, whereas clients who continued in treatment showed a pre- dominance of rapprochement themes (Homer & Diamond, 1996). Hilsenroth, Handler, Toman, and Padawer (1995), using the Minnestoa Multi- phasic Personality Inventory-2 and the Rorschach, examined 97 clients who prematurely terminated psychotherapy and 81 clients who completed at Teast six months of treatment. They found that Rorschach variables of interpersonal relatedness, psychological resources, and level of psychopathol- ogy significantly predicted premature termination. Beckham (1989) found that an initial negative impression of the therapist by the client predicted early dropout from psychotherapy. “There is a sharp contrast between the number of clients who terminate therapy after one session and the attention clinicians give to recommending no treatment for a particular client. With rare ‘exceptions (see Frances & Clarkin, 1981) there has been no research attention given a recommenda- tion of no treatment by the professional asses- sor/therapist following the assessment of clients as the optimal course of action? This discrepancy implies that clinicians almost uniformly recom- mend treatments to those who seek help, while clients often decide after evaluation that pursuing treatment is not needed or indicated, Summary The epidemiological data suggest that only a minority of individuals who need mental health services as indicated by their diagnostic status actually seek assistance from the professionally trained practitioners. If clinicians wish to seek out the many who need psychological assistance but do not seek it, they must make contact with those professionals in the community who come into contact with troubled individuals, for example, physicians, religious leaders, school systems, and divorce lawyers. If the individual does seek assistance, he or she {s almost automatically placed in therapy, with lit-198 + Chapter 6 / The Influence of Client Variables on Psychotherapy i ile clinician attention to those who might handle their difficulties on their own or with watchful fol- lowup. However, many clients, after only one or a few contacts with the mental health system, decide that they can do without assistance. The development of a sharper distinction between those who do leave early and those who follow through on attempts to seek out professional assistance deserves more investigative attention. Among those variables that appear most important are negative attitudes toward the therapist or psy- chosocial treatment in general. In addition, there is ‘an important clinical need to attend to elient rea- sons for prematurely foregoing professional assis- tance from which they could potentially derive some important benefits. Given the large number of clients who leave treatment prematurely, study in this area should be given high priority. PROcEss AND OUTCOME Problem Area/Diagnosis and Severity From a common-sense point of view, all psy- chotherapy should be targeted to the nature of the client$ difficulty, problem, and psychopathol- ogy (depending on one’s conceptualization of the problem area). There should be an inherent match between the clients’ problem area and the therapeutic interventions that are constructed to alleviate or change that difficulty, problem area, and/or diagnostic entity. Diagnosis as the Prescriptive Clem ¥ Variable Following the articulation of DSM-III in 1980 (APA, 1980), this diagnostic template and its sue- cessors have taken center stage in the description of pathology for reimbursement purposes, as well asin planning and guiding psychotherapy research as funded by the National Institute of Mental Health (NIMH). Many cogent arguments can be made for the use of alternatives to a categorical diagnosis, such as dimensional scores on symp- ‘tom and trait measures. However, the DSM sys- tem has guided therapy research, and thus we are accumulating a body of information based on the client variable of diagnosis as defined by the four successive diagnostic manuals, DSM-IV (APA, 1994) defines a mental disor- der as a behavioral or psychological syndrome or pattern that an individual experiences or exhibits as clinically significant because it is associated with distress (¢., a symptom) or disability (e.g, impairment in one or more areas of function or with an increased risk of suffering death, pat disability, or loss of function, In order to facings 4 systematic evaluation of the client with ref ence to mental disorders, general medical con tions, psychosocial and environmental problem and level of functioning, the DSM-IV is @ mule ail system: Axis I—symptom disorders, Aig T—personality disorders, Axis IN—general me. ical conditions, Axis IV—psychosocial and env, ronment problems, and Axis V—a rating of the client overall level of functioning. In actual pag tice, most psychotherapy research is focused og the Axis I condition, with litde research on the Axis II personality disorders. As described later in this chapter, Axis II (personality disorders), IV (psychosocial and environmental problems), and V (overall functioning, related to severity of the illness and impact on functioning) are often empir. ically related to process and outcome of therapy. ‘Much has been written about the advantage and disadvantages of the DSM diagnostic system, ‘The DSM system has been criticized for its pro ‘motion of the medical model to the detriment of biopsychosocial understanding of conditions and their treatments, for its way of defining a mental disorder, for the proliferation of diagnoses acros editions, and for its self-proclaimed atheoretied stance (Nathan, 1998). The conscious meanings of behaviors that are not considered in the DSM criteria are actually most relevant to treatment planning and its execution (Wakefield, 1998) ‘The use of DSM-IV as a guide for pyy- chotherapy outcome research is a mixed blessing. For diagnoses that are closely tied to behaviors such as aleohol and substance abuse, the diagno- sis is tantamount to a description of a problem that is a target for treatment. In contrast, for diagnoses such as depression, there are many routes to such a feeling state, and the behavior that are related to it are often complex and idio- syncratic. From a research point of view, there art problems with selecting a diagnostically “homo- geneous” sample and an appropriate comparison group in order to investigate the impact ofa given intervention, Clients selected solely by the diag- nostic system for a specific disorder are not try “homogeneous” from many points of view. First of all, two clients may actually obtain the samt diagnosis but have very few common symptoms since DSM-IV is polythetic in nature. Second, ‘most clients have more than one diagnosable con- dition or disorder. To use a common clinical situ- ation, twoor its pro- iment ofa itéons and a mental es across heoretca meanings the DSM treatment 998), for py blessing vehaviors, 2 diagno- problem trast, for re many veha and there are “home- parison fagiven he diag ror eruly ow, First he same mptoms Second, ble con cal situ- iteria 0 meet the diagnosis for major depressive disorder, but one client also has one or more Axis II per sonality disorders and the second client has none. Finally, clients with the same diagnosis at best have the same symptoms on either Axis I or Axis I, but other client variables can be quite hetero- geneous. For example, two clients may have exactly the same symptoms that qualify for a major depressive disorder but one is married with a suc- cessful career and the other is unmarried with a poor or absent work history. ‘Thus, the movement to publicize lists of sin- ale DSM diagnoses with empirically supported or validated treatments (Chambless et al,, 1998) can be extremely oversimplified and potentially mis- leading. The lists provide a simplistic algorithm for matching a client witha single diagnosis toa treat- ment for that diagnosis. Such an approach totally ignores the clinical reality that no two clients with the same single diagnosis are truly alike, and these differences are often relevant to treatment planning. Nondiagnostic client variables are totally ‘ignored in ths simplistic approach. ‘We do not review here the extensive research ‘on psychotherapy outcome by the client variable of DSM diagnosis, for this research is extensively covered by other authors in the following chap- ters in this Handbook. Rather, we provide a review of the salient client diagnoses and problem areas that are related to treatment outcome studies There have been a number of reviews of client diagnoses as a characteristic or condition of the client which provides a target for particular types of treatments. The reviews of this literature are ‘rowing, including reviews for government and Practitioners (e.g., Roth & Fonagy, 1996), the {Seneration of treatment guidelines by researchers (Beudler, Clarkin, & Bongar, 2000), independent institutes such as the Cochrane Institute, guilds such as the American Psychiatric Association, and the recent excellent review by the British Psycho- logical Society Centers for Outcomes Research ind Effectiveness for the UK Department of Health (2001). The British review includes the ine reviews in its purview and provides an ‘prto-date summary for client diagnoses and Problem areas including depression, anxiety dis- orders, eating disorders, somatic complaints, per- ‘onality disorders, and deliberate self-harm. Nondiagnostic Client Variables Related 10 Specific Diagnoses With the growing list of psychotherapies that have shown efficacy in the treatment of « specific Process and Outcome » 199 diagnosis as compared to a no-treatment control, some attention has been given to the nondiagnos- tic client characteristics that are related to the process and outcome in these studies. This infor- mation is most abundant as related to depression and substance abuse. For example, Whisman (1993) has reviewed the mediators and moderators of change in the cognitive treatment of depression. Certain key client variables related to the depressive condi- tion have been found to be.mediators of treat ‘ment response; that is, chey mediate the influence of independent variables on the dependent vari- ables in the treatment. The strongest support for ‘mediation was found for attributional style and to a lesser extent for dysfunctional attitudes. There is also evidence that certain client characteristics have a moderating influence on cognitive treat- ‘ments. Sociodemographic characteristics are typ- ically related to outcome (Dobson, 1989; Jarrete, Eaves, Grannemann, & Rush, 1991), whereas intelligence is not (Haaga, DeRubeis, Stewart, & Beck, 1991). Client-learned resourcefulness was related to outcome in one study but not repli- cated in three other studies (Beckham, 1989; Jar~ fett, Giles, Guillon, & Rush, 1991; Kavanagh & Wilson, 1989). A positive outcome from CT was observed in those clients who exhibited a positive expectation of help (Gaston, Marmar, Gallagher, & Thompson, 1989), a strong commitment to treatment (Marmar, Gaston, Gallagher, & Thomp- son 1989), a strong endorsement of the cognitive conceptualization of depression, and a willingness to learn new coping strategies and complete homework assignments. ‘Thase et al. (1997) have taken the research ‘on client variables to new levels by investigating how the sleep profiles of patients with recurrent major depressive disorder are influenced by inter- personal therapy. Those clients with abnormal sleep profiles had significantly poorer clinical ‘outcomes than those with normal sleep profiles, In addition, 75% of those clients who did not respond to IPT manifested remission during sub- sequent pharmacotherapy. Severity of Symptoms Previous reviews of general outcome research have concluded that severity of symptoms is related to poor treatment response (Beckham, 1989; Beutler & Hamblin, 1986; Garfield, 1994, Hoberman, Lewinsohn, & Tilson, 1988; Lambert & Anderson, 1996; Luborsky, Crits-Christoph, Mintz, & Auerbach, 1988). For example, random ow200 + Chapter 6/The Influence of Client Variables on. Psychotherapy regression models were used to examine the role of depression severity in the NIMH Treatment of Depression Collaborative Research Program (TDCRP) Elkin et al., 1995). In this large N, multisite study, the initial severity of depression and the impairment in functioning significencly predicted differential treatment response. There ‘were no differential treatment responses with the less severely il clients, but among those who were ‘more severely depressed and incapacitated, med- ication played a more significant role in combina- tion with psychosocial treatment. Ina study of 117 depressed clients stratified for depression severity (Shapiro et al., 1994), clients were treated in either cognitive behavioral (or psychodynamic interpersonal therapy for either 8 oF 16 sessions. On most measures of outeome, both treatments were equally effective across the severity of depression levels. However, those with ore severe depression improved substantially ‘more with the 16- in contrast to the 8-session ‘weatment duration. Similarly, in the treatment of clients with addictions, those with less severe symptoms demonstrated the best treatment response (MeLel- ‘an, Luborsky, Woody, Druley, & O'Brien, 1983), ‘The six-month treatment outcome for 649 clients ‘who were dependent on opiates, alcohol, and/or Cocaine was examined across 22 treatment settings (McLellan et al,, 1994). Greater substance use at followup, regardless of the abused substance, was predicted by a greater severity of the alcohol and drug use problem at admission to treatment. The severity of the problem, not the number of serv- ices, was the sole predictor of this outcome. In addition, better social adjustment outcome at fol- lowup was negatively related to more severe pay- chiatric problems, employment difficulties, ‘and family problems at admission, Functional Impairment For conceptual clarity and assessment focus, it is important to distinguish between the severity of the symptoms, the major focus of Axis I diag- noses, and the functional impairment that either results from or preceded the symptoms and pro- Fides the context for the arousal of symptoms, Functional impairment is addressed in DSM-IV (on the axis related to overall level of functioning, ‘Two individuals can have a depression of minor Severity or major severity, in the context of pre- vious high-level functioning (productive work, satisfying interpersonal relations) or previous low-level functioning. In general, level of functional impuin negatively correlated with prognosis across fe ders such as depression (Gitlin, Swendsen, Haye & Hammen, 1995; Kocsis eval, 1988; Sony 1991), bulimia nervosa Fahy & Roseel, 9 obsessive compulsive disorder (Keijsers, Hote duin, & Schaap, 1994) and chemical dependeng: (McLellan, Woody, Luborsky, O'Brien, & Deans 1983), In the treatment of depresed nds y the best predictor of response to interp, Psychotherapy (IPT) was emotional health wr to the initiation of weatment Rounswile, Wer man, & Prusof, 1981) Luborsky and collages Gaborsly, 1962; Luborsy etal, 1980) fos significant positive correlation between poy ti: logical health as rated on the Health-Sctny Rating Scale (HSRS) and treatment outcome, nt study of 59 clients treated for 12 weeks wit brief focal psychodynamic therapy, clients whe hhad shown the highest level of adaptive fine tioning before therapy demonstrated the may improvement (Free, Green, Grace, Chernus & Whitman, 1985), In yet another post-hoc analysis of client pre. dictors of treatment outcome for the NIMH s cof the treatment of depression, Sotsky etal (19) examined the treatment of 239 outpatients vik major depressive disorder ina 16-week treatmene, Six client characteristics predicted outcome aeros all treatments (interpersonal psychotherapy, cp. nitive behavioral therapy, medication and clniel management, oF placebo and clinical manage ment), and this included elient dysfunction (oeal and cognitive), expectation of improvement, nd three aspect ofthe symptoms endogenous depres sion, double depression, and duration of curren episode). In addition to these six client character istics which predicted across the treatments, ther ‘were some significant client predictors of « good match with a particular treatment. These authors reported on four such significant matches. Low social dysfunction was a predictor of superior Fesponse to IPT: Low cognitive dysfunction pre dicted response to CBT and to imipramine. High work dysfunction predicted the response © imipramine, and finally, high depression severity and impairment of function predicted response 19 imipramine and to interpersonal psychotherapy. ‘These findings suggest that the focus of the inter vention relates to outcome (eg., low social dye function responses to IPT, which focuses on socil interactions) and that the severity of the condition (symptoms and functioning) calls for a combina- ‘ion of medication and psychotherapy.lent ig iss. gues an) vith og od or Comorbidity "i ‘The pervasive use of DSM-IIl anid its successors in psychotherapy research has fostered examina- ton of the so-called comorbid conditions as they relate to the psychotherapy process and the out- come of a specific symptom-based disorder (see Kendall & Clarkin, 1992). With the distinction since DSM-II (APA, 1980) between symptom conditions (Axis D) and personality disorders (Axis 4), an empirical literature has accumulated con- ceroing the influence of the personality disorders in the treatment of symptom conditions. PERSONALITY DISORDER AND DEPRESSION. Most eudies of major depressive disorder that have included clients with comorbid personality disor- ders have found poorer outcomes associated with co-occurrence of any personality disorder (Burns & Nolen-Hoeksema, 1992; Diguer, Barber, & Luborsky, 1993; Fiorot, Boswell, & Murray, 1990; Greenberg, Craighead, Evans, & Craighead, 1995; Hardy et al., 1995; Shea et al., 1990; Thompson, Gallagher, & Czirr, 1988). The importance of per~ sonality disorder, as a client variable, is also sug- gested by the fact that studies show the reported frequency of personality disorder diagnosis within a depressed population ranges from 24% (Hardy etal, 1995) to 87% (Friedman, Aronoff, Clarkin, Com, & Hurt, 1983). Burns and Nolen-Hoeksema (1992) con- ducted a naturalistic trial of cognitive behavioral therapy for depressed clients and found that bor- derline personality disorder, in particular, was related to poorer outcome. The diagnosis of a personality disorder was related to treatment out- ‘come in the TDCRP study (Shea et al., 1990). Seventy-four percent of the depressed sample in the TDCRP study had a comorbid personality disorder, Clients with personality disorders had ‘Significantly worse outcome in social functioning than clients without personality disorders, and they were more likely to have residual symptoms Of depression at termination. Ina study of 25 clients with major depression Uweated with 16 sessions of supportive-expressive dymamic therapy, clients with personality disor- showed poorer outcome compared to those without personality disorders Wiguer, Barber, & Luborsky, 1993). Hardy et al. (1995), in a tandomized controlled trial of 114 depressed out- Patients seen in either brief psychodynamic inter- Personal therapy (BPI) or cognitive behavioral therapy (CBT), found that the presence of a clus- ter C (anxious-fearful) personality disorder Process and Outcome * 201 reduced the effectiveness of BPI, but not CBT. Finally, others (Fiorot, Boswell, & Murray, 1990; Thompson, Gallagher, & Czirr, 1988) have reported that treatment trials with depressed eld- erly outpatients using behavioral, dynamic, or eclectic therapies have poorer outcomes for clients with a comorbid personality disorder. In a review of 27 different studies, McDer- mut & Zimmerman (1998) concluded that depressed individuals without a comorbid person- ality disorder responded differently to treatment than depressed individuals with a personality dis- order, the latter being more likely to not recover ‘and to remain more symptomatic after treatment. ‘This difference between those symptomatic indi- viduals with and without personality disorder has direct relevance to both the need for an inital assessment and treatment planning (Clarkin & ‘Abrams, 1998). Clearly, the evidence to date sug- {gests that personality disorder, particularly bor- derline personality disorder, is a prevalent and powerful client characteristic that moderates out- come in depressed individuals (Wells, Burnam, Rogers, Hays, & Camp, 1992). Individuals diag- nosed. with borderline personality disorder (BPD) or obsessive compulsive personality disor- der (OCD) have relatively high levels of negative ‘outcome (Mohr, 1995). PERSONALITY DISORDER AS A MODERATOR OF OUTCOME IN ANXIETY DISORDERS. In a sample of 13 outpatients with social phobia, Tamer (1987) found that personality disorder diagnosis predicted differential outcome. Schizotypal, bor- derline, and avoidant personality disorders were related to poor outcome, whereas histrionic and dependent personality disorders were related to better outcome. Clients with dependent person- ality disorders specifically responded better when in-vivo exposure was controlled by the therapist. Studies of anxiety disorders with comorbid avoidant personality disorder have found confice- ing results (Brown, Heimberg, & Juster, 1995, Chambless, ‘ran, & Glass, 1997; Turner, 1987). ‘The presence of a personality disorder has been found to be an obstacle to the treatment of obses- sive-compulsive disorder (AuBuchon & Malat- esta, 1994; Cottraux, Messy, Marks, Mollard, & Bouvard, 1993; Jenike, 1990). AuBuchon and ‘Malatesta (1994) found that obsessive-compulsive clients with comorbid personality disorders responded less well to comprehensive behavior therapy than those without personality disorders. Hermesh, Shahar, and Muniz (1987) found that202 + Chapter 6 / The Influence of Client Variables on Psychotherapy all eight of their borderline clients failed to com- ply with behavioral or pharmacological treatments for OCD. Similarly, Jenike, Baer, Moinichiello, and Carey (1986) found that only 7% of clients with schizotypal personality disorder responded to behavioral treatment, compared to 90% of clients without. PERSONALITY DISORDER AND EATING DISOR- ERS. A number of studies suggest that a comor- bid personality disorder also has deleterious effects on the treatment outcome of eating disor- ders. Rossiter, Agras, Telch, and Schneider (1993) found that eating-disordered patients with comor- bid personality disorders have poor outcome in comparison to eating-disordered patients with- ‘out personality disorders. Cooper and colleagues (Coker, Vize, Wade, & Cooper, 1993; Cooper, Coker, & Fleming, 1994) found that comorbid personality disorder resulted in poor outcome in the treatment of eating disorders. Wilfey et al. (2000), in a randomized controlled study (group cognitive behavioral therapy versus group inter- personal psychotherapy) of 162 outpatients, found that the presence of any Axis II psychopathology did not predict treatment outcome. However, the presence of Cluster B personality disorders did predict poor outcome at one year following treatment. The association found between Axis I disorders and baseline eating-related pgycho- pathology also suggested that this symptomatol- ogy may be more severe when occurring in the context of a personality disorder. This may be because individuals with personality disorders are often likely to have experiences (eg., affec- tive instability, social isolation) that trigger binge episodes. This line of reasoning would suggest that binge eating disorder clients with Cluster B personality disorders may require a specialized treatment that addresses cognitive and affective instability. -HisTony oF SEXUAL ABUSE. Gleaves and Eberenz (1993), in a review of 464 women, assessed the history of sexual abuse in bulimic women who failed to engage in CBT treatment. Approxi- mately 71% of the women who failed to respond to treatment reported a history of sexual abuse. ‘The researchers propose that treatment should address both the eating disorder and the post- traumatic condition, if symptoms and histories of the trauma arise during treatment sessions in individuals failing to respond to CBT. Therefore, although CBT has consistently and convincingly been found to be effective in treating bulimia nervosa, certain client characteristics limie ef’ tive outcome. Summary Use of the DSM diagnostic system to guide py chotherapy research has had both negative ang, positive effects. The emphasis on client diagnos,’ hhas resulted in the lack of attention to oth’ salient client variables (Pilkonis & Krause, 1999, In contrast, a benefit of the DSM multiaxial 5‘ tem is the inclusion of separate diagnostic ae including one for personality disorders, whic have resulted in the accumulation of data on the client variable of personality/personality disorder in the treatment of common symptom disorders" This research has demonstrated across a numbey of symptom disorders that the treatment effecy for the symptoms are attenuated for those client with co-occurring personality pathology in con. trast to those without. This is an important find. ing that should influence treatment planning and facure research efforts aimed at discovering more effective ways of treating those with concurrent symptoms and personality difficulties. At the vey least, therapists should assess for both symptoms and personality disorders in their clinical evalua. tions. When an Axis IT personality disorder is present, they should plan treatment for more modest’ gains, anticipate and address potental early patient dropout, and plan for disruptions in the treatment adherence and alliance. Many of the treatment manuals for symptom disorders such as anxiety and depression give insufficient information on approaches to patients with per- sonality disorders who will present unique and difficult challenges in the treatment. Sociodemographic Variables Age ‘The usual approach to the influence of age on psychotherapy is to assess the relationship in 4 treated group of adults with a limited age range. Te would appear that age is not important in either therapy retention (Berrigan & Garfield, 1981; Dubrin & Zastowny, 1988; Gunderson, et al., 1989; Sledge, Moras, Hartley, & Levine, 1990) or treatment outcome (MacDonald, 1994; ‘Smith, Glass, & Miller, 1980). One possible excep- tion is the finding that younger age is associated with poor retention and outcome in the treat- ‘ment of substance abuse disorders (Agosti, Nunes, ‘& Ocepeck-Welikson, 1996). This latter finding may be due to the relationship between age and the natural courst of substance abuse. thet Met fe place the for ¢ sonal show thea sin Fran tal nig for | ther dlepr rates sabi Tho 19 93: rem resp atk eff eta 198 cog whe chr Aisa ued me coneffec. toms valua- der is ential ny of ders cient per e and ge on ange. nt in field, erson evine, 19% cep cated janes, nding e and A different approach that is currently receiv- ing more attention is to regatd the client's age as an important variable in gauging the focus and ature of intervention. This approach is based on the notion that age is related to the psychologi- cal and biological nature of the organism, and thus to the expression or manifestation of the disorder in question. For example, clinicians intervene with children and adolescents in treat- iments that are structured differently from those for adults. ‘A meta-analysis of 17 empirical studies of the treatment of depressed elderly (Scogin & ‘McElreath, 1994) indicated that psychosocial interventions are quite effective, with a mean effect size of treatment versus no treatment or plicebo of .78, This figure compares well with the mean effect size for psychosocial treatments for depression in nonelderly adults. Interper- sonal psychotherapy, in particular, has been shown to be effective with the elderly in both the acute and maintenance treatment of depres sion in the elderly (Frank et al., 1993; Reynolds, Frank, Houck, & Mazumdar, 1997; Reynolds etal, 1999). Thompson, Gallagher, and Breck- cnridge (1987) have provided empirical support for the effectiveness of cognitive behavioral therapies delivered in the individual format for depression in older adults. In comparing cogni- tive, behavioral, and brief psychodynamic treat- ments for depression in ambulatory elderly, this research group found comparable remission ‘ates across treatment types and no difference in stability of effects for over two years (Gallagher- ‘Thompson, Hanley-Peterson, & Thompson, 1990; Thompson, Gallagher, & Breckenridge, 1987), Although the majority of clients achieved Temission, a subgroup of clients who did not Tespond to initial treatment, remained depressed st followup one and two years later despite eon tinued treatment. Cognitive behavioral treatments are also fective when delivered in a group format (Arean stal., 1993; Beutler et al., 1987; Steuer et al., 1984). Kemp, Corgiat, and Gill (1992) found that Cognitive behavioral group therapy was effective in teducing depressive symptoms in older clients ‘ho had the presence or absence of disabling ‘hronic illness. In contrast, however, those with dissbling physical illnesses did not show contin- ved decline in depression following group treat- ment, while those without disabling illnesses Continued to improve. bi Process and Outcome * 203 Socioeconomic Status In general, demographic characteristics and socioeconomic status (SES) have been found to be related to continuation in psychotherapy. Early studies (Berrigan & Garfield, 1981; Dodd, 1970; Fiester & Rudestam, 1975) found a positive relationship between higher social status and length of stay in treatment. For example, Arm- buster and Fallon (1994) found lower SES to be associated with premature termination among ‘general psychotherapy clients. In the treatment of substance use, a shorter length of stay was associ- ated with lower educational background (Agosti, Nunes, & Ocepeck-Welikson, 1996; Epstein, ‘McCrady, Miller, & Steinberg, 1994; McCusker, 1995). These results are not always consistent, however, and one can also find other studies in which SES was not related to terminating or remaining in treatment (e.g., MacDonald, 1994; Sledge, Moras, Hartley, & Levine, 1990). Gender Prior reviews make the generalization that there is usually no gender difference in premature termi- natién from therapy or any gender effects in psy- chotherapy outcomes (Garfield, 1994; Greenspan & Kulish, 1985; Petry, Tennen, & Affleck, 2000; Sledge, Moras, Hartley, & Levine, 1990). In con- ‘trast to gender effects in general, gender might make a substantial difference with disorders that hhave a prevalence rate that is related to gender, such as depression. The prevalence of depression is about twofold in females in comparison to males (Kessler, McGonagle, Swartz, Blazer, & Nelson, 1993; Weissman & Klerman, 1977). In addition, the cause of depression may be different in females in contrast to males (Cyranowski, Frank, ‘Young, & Shear, 2000; Nolen-Hoeksema, 1987). Despite the differences in prevalence and causes, with few exceptions, sex has been unrelated to ‘outcome in the treatment of depression (eg., Hol- Jon et al, 1992; Paykel et al, 1999; Sotsky et al., 1991), Unfortunately, studies are rarely designed specifically to study this issue, and occasionally there is an exception to the lack of difference due to gender. For example, Thase, Frank, Kornstein, and Yonkers (2000) found across studies that ‘women who were manifesting more severe depres- sion did better in interpersonal therapy than they did in cognitive therapy. This result sug- gests that the search for gender differences in interaction with treatment is worth pursuing in farure research.204 * Chapter 6 / The Influence of Client Variables on Psychotherapy ‘There has been some attention to same-sex pairing between client and therapist, with some finding same-gender pairing providing greater client satishetion and retention in treatment ujino, Okazaki, & Young, 1994) and others showing preference for opposite-gender matches (Willer & Miller, 1978). One large study Flaskerud & Liu, 1991) found that client-therapist gender similarity had livde effect on outcome. These inconsistent results suggest that the more sophis- ticated methods may reveal some advantage to matching and that further testing for matching within specific problem areas may reveal opti- ‘mal matches. Race Several early studies found that ethnic minority clients attended significantly fewer sessions than Caucasian clients (Greenspan & Kulish, 1985; Saleman, Shader, Scott, & Binstock, 1970; Sue, ‘McKinney, Allen, & Hall, 1974). However, other studies found no relation between race and pre- mature termination (Sledge, Moras, Hardey, & Levine, 1990). Well-controlled research by Jones (1978; Jones & Zoppel, 1982) found that race- related client and therapist variables (eg, race matching between therapist and client) were not decisive in therapy outcome. African-American and Caucasian clients benefited equally, and no differences were found between racially matched ‘or mismatched therapist-client dyads. Lemer (1972) investigated the effects of treatment on severely disturbed and predominantly lower class Aftican-American and Caucasian clients seen by Caucasian therapists. The vast majority of clients improved, and there was no evidence of racial dif- ferences in outcome. In addition, she found that low-income clients, regardless of race, showed ‘more improvement in therapy when seen by ther- apists holding egalitarian attitudes toward low- income people in general than did clients not seen by therapist holding such attitudes. Ross (1983), using the same measure of therapist atti- tude, found that low-income African-American clients remained int therapy longer when seen by therapists with egalitarian attitudes. Within such a context, as the client communicates both ver- bally and nonverbally, the therapist allows himself or herself to empathize with the client’ emo- tional position and develop an involving intersub- jective perspective with the client. Thus, Lerner (1972) and Ross's (1983) research on the impact of therapist attitudes on treatment outcome found that egalitarian atmosphere is an important variable in work with lower-class client, studies by Lerner and Jones represent some of most detailed and rigorous treatment s involving African-American clients. ‘Occasionally, race-based differences are and suggest the need for continued research. Fy ‘4 ‘example, Rosenheck, Fontana, and Cottrol (I po eH found that African-American veterans with a thee di traumatic sess disorder were more ikely odes srs cout of therapy and were less likely to benefitfon’ff rere is treatment than their Caucasian counterpart jexsuse Unfortunately, these researchers did not stu therapist ethnic group identification. YE Summ Various writers note how therapists talk of the inf “properly managing” the initial sessions eatcom clients of color (Griffith & Jones, 1979; Js seany € 1997; Sue & Zane, 1987). Griffith and Tost vox ch (1979) have suggested that effective work wid sheeap African-American clients, especially when the axnwnpl therapist is Caucasian, involves working quia aiid eat to establish a therapeutic alliance. Jenkins (997) sera points out that it is important to emphasize te} hav bee quality of the therapist-client relationship as fun! tel (0 damental to positive change, especially whea ‘von pr working with ethnic minority clients. Sue and. swore d Zane (1987) note the importance of the thers’ peutic pists establishing his or her “credibility” early on D ‘with the ethnic minority client. Gibbs (1985) sug. vw thet gests that African-American clients, mindful of-f} slated racism, initially tend to take an interpersonal of any entation in the therapy situation. That is, they are Jest particularly sensitive to the process going on wpmen ‘between themselves and their therapists. Sue and during Zane (1987) contend that ethnic-minority clients inc come to believe in the credibility of therapiss through two factors: ascribed and achieved statis » era “Ascribed status is the position or role that one is own id assigned by others, usually based on factors such . eee as age, expertise, and sex. Achieved credibility seckin refers more directly to therapists’ skills. Through career the actions of therapists, clients come to have satsta i t, confidence, or hope. Unfortunately ¢ fea ment benefits i seldom studied and remains f * largely untested. and c In addition to the questions relating to the © thar race/culkural background of therapist and patiens | aa there are potential research questions concerning me the relationship between race and the nature, fee , fe ; tures, expression ofthe problem area, or diagnos: FFF tic issues faced by the client. For example, the ae presence of eating disorders in Caucasian and © ca African-American women has been found t0 naa ‘manifest a different pattern of pathology (Pike1 Dohm, Stiegel-Moore, Wilfley, & Fairburn, 2001). In both groups, eating disorders are asso- Gated with decrements in functioning, but the ‘wo groups differ on aspects of the eating disor- der, including binge frequency, restraint, treat- ment-seeking behavior, and personal concerns about eating, body weight, and shape. Although these differences suggest that differential actions by therapists might result in different outcomes, there is, as yet, no evidence that outcomes vary becease of these differences. Sunimary The influence of client demographic variables on coutrame is mixed and inconsistent, possibly for mang reasons. Attitudes toward age, gender, and race change with the times, and both patients and therapists will be influenced by the cultural atmosphere. The clients age, gender, ethnicity, and education are fixed variables to which the therapist must accommodate and adjust. There has been a growing recognition of the need in the feld for training therapists in this accommoda- tion process, and this ability is probably learned more directly from supervisors than from thera- peutic manuals. Demographic variables may be less important in themselves and are rather a marker for other related issues. For example, age is a marker for ‘many aspects of clients’ lives that are relevant to tweatment planning. Age correlates with the devel- ‘pment of the biological organism that unfolds during childhood to adolescence and declines dur- ing old age. Age correlates with the developmen tal and psychological tasks that an individual faces over life span. Adolescents are establishing their own identity and making moves toward independ- nce from the family of origin. Young adults are seeking intimate parmers and beginning work ‘steers. Middle-age adults are facing the tasks of ‘scisfaction in intimacy and raising children Advancing age brings issues of health, loss of red ones, and diminishing activities. Our society is attentive to issues of ethnic and cultural diversity. Icis commonly taught now that the clinician must be attentive to the ethnic 4nd cultural aspects of the client in order to form 2 fruitful therapeutic relationship. This orienta~ ton can be taken to an extreme form by calling replication of all psychotherapy studies by agnosis (EST literature) with all ethnic groups all, 2001). This argument would be more con- Ne if there were signs that ethnic differences a significant influence on treatment outcome cr that ethnic variables were related to the nature of the conditions being treated. ‘The most fruitful areas of future research involve those in which the nature and manifesta- tions of the problem area or diagnosis are related to the demographic characteristics of the clients. ‘The two best examples are those reviewed in this section relating to depression and eating disor- ders. The prevalence rate and the experience of depression are related to gender. Race may have an influence on the pattern of eating disorders. Fature research is needed to examine not only the treatment prognosis but also the issue of prescrip- tive treatments for depression as related to gender. Personality Variables ‘Under the heading of diagnosis, we have previ- ously considered the influence of personality dis- orders as defined in DSM on psychotherapy. ‘Most reviewers consider the personality disorders to be an extreme of personality traits, with conti- nuity between normality and disorders. In this section, we consider other personality traits as they influence the therapeutic encounter. Expectancies ‘There is a history of research relating client expectancies and therapy process and outcome (Frank, 1973), Paul and Shannon's (1966) work on systematic desensitization) found that a positive expectancy condition yielded a better outcome than a no-treatment control. Frank (1961) consid- cred the clients’ confidence in his or her therapist and treatment to be the critical determinant of ‘outcome. Client expectations of treatment were related to treatment duration (Lorr & McNair, 1964), attrition rates (Overall & Aronson, 1963), and outcome (Lennard & Bernstein, 1960). Gas- ton, Marmar, Gallagher, and Thompson (1989) found better outcomes for cognitive therapy clients who expected the treatment to work. Client expectancies have a strong relationship to duration of treatment (lenkins, Fuqua, & Blum, 1986) but an inconsistent relationship to treat- ‘ment outcome (Beutler, Wakefield, & Williams, 1994), In a study of brief ambulatory psychother- apy Joyce & Piper, 1998), client expectancies were associated strongly with the treatment alliance but only moderately related to treatment outcome. In the same study, client expectancy and a measure of quality of object relations combined in an additive fashion to relate to both alliance and outcome. There is evidence that client expectancies and “difficulty” are related to therapist behavior206 Foley, O'Malley, Rounsavill, Prusoff, & Weiss- ‘man, 1987) in delivering a manualized IPT treat ment. Client difficulty as demonstrated in the therapy sessions was related to therapists’ and Supervisors’ judgments of therapist performance; that is, therapists were seen as performing more Poorly when clients were more difficul. Clients’ Pretreatment negative expectations about the Outcome of therapy were associated with client difficulty, whereas level of presenting symptoma- tology was not. Preparation for Change A number of constructs describe the client's own preparation for behavioral, attitudinal, and emotional change as it intersects with help- seeking behavior, READINESS To ctuaNGE. When the client makes 8 decision to seek therapy as a means of dealing With difficulties, to what extents the client moti= vated to do what is necessary for change? Prior to coming for therapy, what efforts has the client made to make changes in order to overcome his or her difficulties? These basic questions have been examined extensively in relation to the issue of terminating the habitual and harmful behavior of smoking. DiClemente and Prochaska (1982),described and assessed the frequency of 10 change processes in individuals who smoke. From this data set these investigators (Prochaska & DiClemente, 1983) described a series of five stages in the ces” sation of smoking: (1) precontemplation in which People are not intent on taking action, (2) con- templation in which people intend to take action, @) Preparation in which people intend to take immediate action, and finaly (4) an action stage in which individuals make specific modifications in their behavior and (5) maintenance in which individuals take steps to avoid relapse to the undesired behaviors, Addiction severity and frequency of smok- ing per dag were significantly lower among those in the preparation stage than those in the pre- contemplation or contemplation stage (Crittendon, Manfredi, Lacey, Warnecke, & Parsons, 1994, DiClemente et al, 1991). During an intervention study, clients in the preparation stage made greater use of the intervention (as predicted) than did Precontemplators or contemplators (DiClemente etal, 1991), In yee another study (Farkas etal, 1590), clients in the preparation stage were more likely to have stopped smoking one to two years * Chapter 6 / The Influence of Client Variables on Psychotherapy later compared to clients inthe contemplate precontemplation stages. ‘ ‘The stages of change have been app seven ferent systems of psychotherapy (a chaska & DiClemente, 1983, 1984, 1983) "%% dropout rate from treatment for a varieny 97 orders such as substance abuse, smoking chat and medication treatment for hypertension’ HIVIAIDS has been related to stages of oh Medeiros, Prochaska, & Prochaska, in preg Prochaska, Norcross, Fowler, Follick, & Ajo % 1992), Sagecelated variables were moze pos fal than demographic variables, type and sever of problems, and other client variables. Funded more this group has made the prediction that amount of change during treatment and fol ing treatment is significantly related to the stage of change at the beginning of treatment (pt chaska, DiClemente, & Norcross, 1992) A chy cal corollary or principle stated by this group that the treatment should be matched to de clients stage of change and that 2 mismaxd becween client stage and therapist strategies nj result in resistance, In a large clinical trial, four treatmeny were compared for 739 smokers (Prochasl DiClemente, Velicer, & Rossi, 1993). The fox treatments included a home-based cessation pro ‘gram, 2 stage-matched individual treatment, an expert system computer report plus manualized eamment, and fnally counselors plus computer and manualized treatment. At 18 months the stage-based and matched programs were superior to the other treatments. Results are not always consistent, however. For example, Ziedonis and ‘Trudeau (1997) evaluated stage of change among a large group of community mental health centet clients with schizophrenia spectrum diagno:tt and substance use disorders. Their results did net support the validity ofthe predictions conceming stage of change and involvement in substance abuse treatment or its outcomes. It appears tut the stage strategy often is predictive and can be used to design interventions, but the majority of research is on habit disorders and must be inves- tigated in other client problem areas. Ego Strength An important factor known to affect treatment outcome is clients’ ego strength (Kernberg et aly 1972; Sexton, Fornes, Kruger, Grendahl, & Kolseth, 1990; Sohlberg & Norring, 1989). Ego strength is defihed as the presence of positive per- sonality assets that enable an individual to toler-lation o¢ axe and overcome his or her anxieties and to acquire new, more adequate defenses. According plied to 10 Brown (1979, p. 184), “Egh strength is also the °¥ Pro. client’ capacity to hold on to his own identity 3). The despite psychic pain, distress, urmoil and conflict 1 of dis berween opposing internal forces 2s well a the obesity, demands of reality.” Consistent with these defini. ion and ‘ions, research has shown that those scoring high change on ego strength measures are rated as better 1 press, adjusted psychologically and show a greater Abrams, capacity to cope with the stressors and problems Power. in their life situations (Graham, 1990). Ego severity strength and similar concepts have also been gen ‘urther- erally found to be related to treatment outcome that the (Conte, Plutchik, Picard, & Karasu, 1991; Kern- follow. berg et al, 1972; Sexton, Fornes, Kruger, Gren- ve stage dahl, & Kolseth, 1990; Sohlberg & Norring, © Pro 1989) In the Menninger Psychotherapy Project, Acdini- Kemberg et al. (1972) found a significant rela. roup is donship between ego strength and outcome in to the sychoanalytically oriented psychotherapy (R[df= smatch 41] = 35, p < 0). Exceptions to these findings ies will include studies by Luborsky et al. (1980) and Endicott and Endicott (1964), both of whom: tments found that the Barron's Ego Strength Scale was chaska, not significantly related to outcome. In addition, xe four in the Columbia Psychoanalytic Center, Project, on pro- clinical appraisals of ego strength were not signif nt, an icantly related to outcome (Weber, Bachrach, & alized Solomon, 1985), Whether ego strength influ- puter frees particular aspects of the therapy process hs the (ie., formation of the therapeutic alliance and ‘perior | silty to obtain insight) or exerts direct effects on ate ‘outcome is in need of further exploration, is an among Prychological Mindedness center McCallum and Piper (1996) have reviewed the ‘mnoses client construct of psychological mindedness, in lid not terms of its definition, assessment, and relation- ering ship to outcome. Psychological mindedness (PM) ‘stance ‘fers to a person’ ability to understand people ‘3 that 2nd their problems in psychological terms: From aan be | 4 psychodynamic Perspective, PM refers to the city of ability to identify ‘components of intrapsychic inves- conflict. There are self-report measures of psy- ‘ological mindedness, including @ subscale of girh a self-consciousness scale (Fenigstein, ing gt & Buss, 1975), and clinical interviews, ment including the Psychological-mindedness assess- etal, "Ment procedure developed by these authors a & Baer, Dunbar, Hamilton, and Beutler (1980) . Ego Sctor analyzed therapist ratings of process items eper it found that a patient’. demonstration of higher toler ‘eels of insight and self. disclonane ‘was related to Process and Outcome © 207 ‘treatment outcome. The Psychotherapy Research Project of the Menninger Foundation failed to find any significant relationships between ratings of psychological mindedness and outcome. In a Comparative outcome study, Piper, Debbane, Bienvenu, and Garant (1984) found that psycho. logical mindedness was significantly related and directly correlated with client outcomes in a short-term group therapy but was not predictive of eutcome in the other three forms of therapy studied (long-term individual therapy, long-term group therapy, and short-term individual). In a controlled trial of an interpretive form of short. term group therapy, PM was directly related to remaining and working in groups but not to deriv. ing benefit from them (Piper, McCallum, & Azim, 1992), In a day treatment trial, however, PM wae directly related to both working and benefiting. ‘The authors hypothesize that clients with higher levels of PM are better able to work and benefit in interpretive therapy in which internal conflicts are explored repeatedly, and, conversely, clients with lower levels of PM are better able to work and benefit in supportive therapy in which intemal conflicts are not explored. Avidiynie-mrozecrive pistincrion. Blatt tal. (1994) found that in long-term treatment, clients who were predominantly introjective (per? fectionistc and self-critical) had generally beeer outcomes than clients who were predominantly anacltic (concerned with abandonment and los). In using the perfectionism subscale of the Dye. functional Ardtude Scale (DAS) as am analogue for introjective syle, it was found that prement, ment perfectionism had a significant negative impact on therapeutic outcome across teats ‘ment conditions Blat, Quinlan, Pilkonis,& Shes, 1995). The distinction between anacltic ancl introjective clients was also applied to a further analysis of data from the Menninger Peychother, apy Research Project (MPRP). Findings indicated that anaclitic and introjective clients are differen: tially responsive to psychotherapy and psycho. analysis. Anaclitc clients had significantly greater improvement in psychotherapy than they did in . In contrast, introjective clients had signifcanty greater improvement in paycho- analysis than in psychotherapy. Interpersonal Variables Interpersonal Relatedness One of the most frequently studied client factors is the clients quality of relating in interpersonal208 + Chapter 6 / The Influence of Client Variables on Psychotherapy relationships (Luborsky, Barber, & Beutler, 1993). Interpersonal relatedness has been conceptualized in a variety of ways by a number of investigators from different theoretical orientations. Some investigators have assessed the history of interper- sonal relationships, whereas other investigators have examined interpersonal functioning in cur- rent close relationships or have assessed cients? perceptions, beliefs, and wishes about relation- ships. Others have looked at how the client relates to the therapist during the therapy. ‘A number of investigators have demonstrated significant relationships between the pattern of a client’ pretherapy interpersonal relationships and the therapeutic alliance established during treat- ment (Luborsky, McLellan, Woody, O'Brien, & Auerbach, 1985; Marmar, Weiss, & Gaston, 1989; Piper, Azim, Joyce, & McCallum, 1991). The find- ings in this area are somewhat mixed. Piper et al. (1991) found that the greater the disturbance between a client and his or her parter, the better the alliance established with the therapist. Those clients who were emotionally needier established longer term relationships, compared with those participants who did not seem to have such needs and stopped treatment prematurely. In addition, disturbance with one's partner is but one aspect of problems in interpersonal relationships, thus ing the client more toward an important positive alliance with the therapist. In contrast, in an. uncontrolled followup study of 84 clients treated with individual psychoanalytic psychotherapy, the capacity to be related was significantly predictive of positive outcome (ClementelJones, Malan, & ‘Traver, 1990). These findings are also consistent with those of Alpher, Perfetto, Henry, and Strupp (1990), who found a significant positive relation- ship between clinician ratings based on clinical interviews of clients’ capacity to engage in short- term dynamic psychotherapy and clinical out- come as assessed on the Rorschach test. Moras and Strupp (1982) also found that good pretreatment interpersonal functioning predicted a good alliance, but they did not find that poor interpersonal func- tioning predicted a poor alliance. On the other hand, ‘consistent with Piper’ findings, Walters, Solomon, and Walden (1982) report that clients who remained in treatment were more poorly adjusted than those who terminated prematurely. Tt is plausible that clients with disturbed interpersonal functioning are nevertheless so dependent and needy of interpersonal relation- ships that they continue to stay in therapy despite problems expressing their needs and difficulties in their personal relationships. Corresponding, * those who drop out prematurely may do s¢/ because they have lower needs for closeness ang intimacy, regardless of whether or no they hae better overall interpersonal relationships or are! beter at denying interpersonal conflicts. Along * these lines, in 2 15-month followup assessment of clients at long-term psychoanalytically oriented treatment facility, Blatt, Ford et al. (1994) found" that those clients who made substantial clini! progress (defined as less frequent or less severe clinical symptoms and more intact social bebay. jot) had produced more disrupted and malevolen, interpersonal interactions on the Rorschach in their initial intake assessment, The authors sug. gested that clients who are more open about their disturbed interpersonal relationships are more likely to enter actively into therapy and to gain most from the treatment process. Quality of Object Relations Interpersonal relatedness has also been conceptu- alized in terms of the quality of object relations, Quality of object relations refers toa person’ life Tong pattern of relationships and their character- istic way of interpreting social information. The quality of object relations is believed to be + dimension ranging from immature to mature le els of relatedness. In a comparative psychother apy study, the therapist’ rating of quality of object relations was directly related to favorable process and outcome in an interpretive form of short-term individual therapy (Piper, de Carufl, & Sakrumelak, 1985). In 2 controlled tral of interpretive, short-term individual therapy, qual- ity of object relations was directly related to the therapeutic alliance and favorable outcome (Piper et al, 1991). These findings are consistent with those reported by Horowitz, Marmar, Weis DeWitt, and Rosenbaum (1984) in a study of brief individual therapy. In addition, in a con- trolled trial of intensive day treatment, which involved an integrated set of interpretive and supe portive forms of group therapy, quality of object relations was directly related to remaining in and benefiting from treatment (Piper, Joyce, Azim, & Rosie, 1994). The authors conclude that cliens : with higher levels of Quality of Object Relations; are better able to tolerate, work with, and benefit, from the more demanding aspects of interpret: therapy, and, conversely, clients with lower of quality of object relations are better able ©, work with and benefit from the more rac aspects of supportive therapy. : Attache since psy} ofa relatio (jes client sherapist)> arrachment drachment rocess a0 Seminal wc Ainsworth infants and ment has behavior. 7 the caregis analogous; ubsress see therapist. sryles has ambivalent (Bowlby, 1 Inat with borde’ chissifed v {AAD as ir response ¢ attachment uaalistic satiety of 2 substance Heape, and ‘ypes follo described b lewas foun trast to var ated with fi of treatmer lowing trea Since | objece refe client’ life these const ment proc immediate cies and el Fesponses fi There is e Individuals Perceive th ships and e artholom W991; Grit (1990) four tesistant toAttachment Patterns Since psychotherapy involves the creation and use ofa relationship between two or more individuals (ie, client and therapist, client, spouse/family and therapist), itis plausible thatthe clients’ history of arachments to others and the quality of these acachments will have a predictive effect on the process and outcome of treatment. Following the seminal work by Bowlby (1969, 1980, 1988) and ‘Ainsworth (1964) on the attachment between infants and their mothers, the construct of attach- ment has been examined in relation to adult behavior. The attachment behaviors of infant to the caregiver under situations of stress may be analogously related to the situation of a client in distress seeking a help-giving relationship with 2 therapist. The nature of attachmentor atachment styles has been described as secure, anxious- ambivalent, anxious-avoidant, and disorganized (Gowiby, 1988). Th a treatment study of clients diagnosed with borderline personality disorder, those clients classified with the Adule Attachment Interview (AAD as insecure-dismissive evidenced the best Tesponse to intervention compared to other attachment groups (Fonagy et al, 1996). Ina nat- uralistic treatment study of outpatients with a variety of Axis I disorders (eg., affective, anxiety, substance abuse), Meyer, Pilkonis, Proietti, Heape, and Egan (2001) rated attachment proto- ‘ypes following an interview using a procedure described by one of the authors (Pilkonis, 1988). Jewas found that secure attachment style in con- trast to various insecure attachments was asso ated with fewer symptoms prior to the initiation of treatment and with greater improvement fol- lowing treatment. Since both attachment style and quality of object refer to relationships with others in the client’ life rather than with the current therapis these constructs may have effects on the tre ment process and/or outcome through more ‘nmediate mechanisms, including client expectan- ies and clicitation of helpful versus harmful ‘sponses from the therapist (Meyer et al., 2001). ere is evidence that this might be the case. Individuals characterized by secure attachment Peteeive themselves to be competent in relation- ips and expect a positive response from others ,lomew, 1997; Bartholomew & Horowitz, 1991; Griffin & Bartholomew, 1994), Dozier (1990) found that dismissing patients are often "sistant to treatment, have difficulty asking for he Process and Outcome © 209 help and retreat from the help that is offered. Dismissing individuals often become disorgan- ized when they are confronted with emotional issues in therapy (Dozier, Lomax, & Tyrrell, 1996). Satterfeld and Lyddon (1998) found that security of attachment was related to positive scores on the goals subscale of the Working Alliance Inventory (WAI, Horvath & Greenberg, 1986). Eames and Roth (2000) found that patient attachment orientation was related to the develop- ment of a therapeutic alliance during tlie early stages of therapy. Attachment status was also related to the frequency of therapeutic ruptures. Security of attachment was related to higher ther- apist-rated alliance, and fearful avoidance was related to lower levels of alliance. Interestingly, the preoccupied attachment dimension was related to low alliance at the beginning of treatment but higher alliance toward the end of treatment. ‘These findings, taken together, suggest that anxi- ety about attachment and avoidance of intimacy may actto impede the development of a therapeu~ tic alliance. However, regardless of high levels of anziety about relationships, the strong drive of highly preoctupied individuals for intimacy might enable them to develop a better alliance as therapy continues. Surprisingly, Eames and Roth (2000) also found that dismissing attachment was related to positive changes in alliance during the course of therapy. Malinckrodt, Gantt, and Coble (1995) also found a subgroup of patients they called relue- tant, who reported good alliances on the WAI but endorsed an unwillingness to participate in the self-revealing tasks of psychotherapy on the Client Attachment to Therapist Scale. ‘These authors suggested that the reluctant cluster might corre- spond to the dismissing category. Patient attachment may also influence alliance by influencing therapist response. Dolan, Amkoff, and Glass (1993) found evidence to sug- gest that therapist and client attachment styles ‘were interdependent and that ratings of working alliance were contingent on perceptions of thera- pist-client differences. Hardy, Stiles, Barkham, and Startup (1998) examined responses to patient attachment patterns and found that therapists tended to adopt more affective and relationship- oriented interventions in response to clients with ‘overinvolved-preoccupied interpersonal styles and used more cognitive interventions with patients characterized a5 underinvolved-dismissing. Patients in treatment with therapists who were dissimilar from them on the hyperactivat-210 + Chapter 6 / The Influence of Client Variables on Psychotherapy ing/deactivating dimension of atachment on the Adult Attachment Interview (AAI showed better therapeutic outcomes and stronger therapeutic alliances (Dozier, Cue, & Bamett, 1994; Tyrell, Dozier, Teague, & Fallot, 1999). Clinicians classi- fied as secure/autonomous on the AAI tended to challenge the patient’ interpersonal style (whether deactivating or hyperactivating), while clinicians classified as insecure on the AAI were more likely to complement the patients’ interpersonal style Dozier et al., 1994; Tyrell, Dozier, Teague, & Fallot, 1999). Patients treated by clinicians classi- fied as secure on the AAI have the best outcomes when the clinician is at the opposite side of the secure/autonomous continuum from the patient’ AAT classification (eg., the patient is rated Preoc- cupied on AAT, and the therapist is rated on the dismissing end of the autonomous category Fl, ¥F2) (Dozier et a., 1994). Diamond and colleagues (Diamond et al. 1999) reported findings from two clients with bor- derline personality disorder treated in Kemnberg’s transference-focused psychotherapy (Clarkin, Yeo- mans, & Kernberg, 1999) by the same therapist. Both clients progressed from insecure to secure states of mind regarding attachment with one year of treatment. However, consistent with previous research Eames & Roth, 2000; Dolan, Amkoff, & Glass, 1993; Dozier et al, 1994; Mallinckrodt, Gantt, & Coble, 1995; Tyrell et al, 1999), each patient interacted and affected the therapist in very different ways, and the therapist responded to each patient very differently. The therapist was engaged and active in the treatment of the client initially classified as preoccupied, whereas the same thera- pist was much less engaged, often felt dismissed, and developed a much weaker therapeutic bond with the other client. In-Therapy Bebavior In many studies, clients’ characteristics are meas- vured with paper and pencil assessment instru- ‘ments or are determined through semistructured interviews. A more direct test of clients’ charac- teristics is to assess the clients’ behavior during the therapy itself, such as their contribution t0 the therapeutic alliance and involvement in the treatment process. Cumvr participation. Gomes-Schwarte (1978) analyzed process ratings from taped segments of therapy sessions and found that the feature most consistently predicting outcome was client will- ingness and ability to become actively involved in the therapy. In addition, O”Malley, Suh, and Sry (1983) found that client involvement correla significantly with all measures of outcome in the Vanderbilt Psychotherapy Outcome Study. Nel son and Borkovec (1989) found that canonicg correlations of participation correlated with change ‘on pre-post outcome measures. THERAPEUTIC ALLIANCE. Client characterising such as the ability to form an alliance with the therapist and initial functioning also. proved important in predicting treatment outcome, Research has indicated that the clients’ contiby. tion to the therapeutic alliance is related to ther. apy outcome (Horowitz, Marmar, Weiss, DeWin, & Rosenbaum, 1984; Marziali, Marmar & Krup. nick, 1981). Krupnick et al. (1996) found thi ‘mean therapeutic alliance, assessed in the third, ninth, and fifteenth sessions, was significandy related to outcome across treatment groups. This relationship was determined primarily by the con tributions of the client rather than by the therapist to the therapeutic alliance. Using the Vanderbilt Psychotherapy Process Scale (VPPS), Windhol: and Silberschatz (1988) found that the cliens’ involvement in the relationship and the therapis- offered relationship were significantly correlated with the therapist’ rating of outcome in a brief psychodynamic therapy. With 86 clients manifesting anxiety, depres sion and personality disorders, alliance signif cantly predicted subsequent change in depression when prior change in depression during the treatment was partialed out (Barber, Connolly, Crits-Christoph, Gladis, & Siqueland, 2000). ‘The authors suggest that their design and find- ings advance the research question in this are from whether therapeutic alliance during the fst few weeks of psychotherapy predicts outcome to the question of the nature of the intertwined and sequential relationship between alliance and clients’ improvement. Summary: Interpersonal Bebavior Psychotherapy involves an interpersonal process between client and therapist. The clients’ pist interpersonal relationships and current ability © form a positive and fruitful relationship with the therapist are, on the face of it, quite relevant the continuation and success of the thertPh This situation is, in some ways, a dilemma, it that many symptomatic individuals with disor ders needing' treatment are the same ones WHO have troubled interpersonal relations that ™¥pp ated the Nel. ried stig ved ome, iby. her. ees pt | ye | athe met apy: a, it igor a may disrupt the therapeutic venture. Research support for the importance of these variablesis abundant in the literature despite differing operationalizations and diverse treatment methods. Therapists must be expers in fostering relationships with individu als who have diffculry doing so. Search for a Set of Client Characteristics Juis quite plausible that single-client variables will rut prove to be as important to the treatment yeess and outcome as a set of interrelated client fariables. Several teams of researchers have fearched in different ways to find sets of client ‘eriables that have implications for outcome. Client Variables across Problem Areas/Diagnoses One of the most systematic and concerted efforts to isolate a set of specific client variables and demonstrate their influence on the course and ‘outcome of treatment has been the work of Beut- ler and his colleagues. Beutler, Clarkin, and Bon- ¢g2t 2000) have recently documented the steps in identifying salient client characteristics that are potentially related to treatment process and out- come. First, comprehensive reviews of treatment studies were utilized to describe client character- istics (Beutler, 1979; Beutler & Berren, 1995; Beut- ler & Clarkin, 1990; Beutler, Consoli, & Williams, 1995; Beutler, Goodrich, Fisher, & Williams, 1999; Beutler, Wakefield, & Williams, 1994; Gaw & Beutler, 1995). Second, based on an extensive list of lent variables, an attempt was made to extract the more trait-like characteristics that might have an ‘enduring impact on the treatment process and out- ‘come across time. This was followed by an attempt torrelate these trat-like client variables to differen- tal aspects of the pharmacological and psychosocial ‘treatments employed. Since this chapter focuses entirely on client variables, we will provide an examination of treatment modifiers that were isolated to optimally match the client and treat- ‘ment interactions. The interested reader can pur- ‘ue a more complex analysis in Beutles, Clarkin, and Bongar (2000) as well as in Chapter 7 of this volume, Finally, Beutler and colleagues developed ‘methods for assessing the client variables and con- ted a predictive validity study using these client fattles to predice treatment outcome Beule, ‘Moleiro, Malik, & Harwood, 2000). for jut st client variables identified and selected ty) estgation out ofa large number of poten ‘al candidates included: (1) client functional Process and Outcome © 211 impairment, (2) subjective distress, (3) social sup- ‘port, (4 problem complexity/chronicity, (5) client reactance/resistance, and (6) coping styles. These ‘variables relate to the client's problems and psy- chopathology (complexity/chronicty, functional impairment), to the characteristic ways in which the individual responds to difficulty (subjective distress, reactance/resistance, coping styles), and to the nature of the client’ interpersonal context Gocial suppor). “Two client variables—functional impairment and complexity/chronicity of problems—relate directly to the clients problems, illness, and/or psychopathology. Complexity may be defined as, comorbidity (e., coexisting diagnosable symp- tom or Axis I disorders and/or coexisting Axis II or personality pathology) and the duration of the difficulties Ge, the chronicity, frequency, and extent of recurrence). There is evidence that greater problem complexity calls for more com- plex and broadband treatment. For example, situ- ation-specific problems, as opposed to chronic and recurrent problems, have been found to be more responsive to behavioral treatments. This seems to be true for those with mixed somatic symptonis (LaCroix, Clarke, Bock, & Doxey, 1986), alcohol abuse (Sheppard, Smith, & Rosen- baum, 1988), eating disorders (Edwin, Anderson, & Rosell, 1988), and chronic back pain (Tief & ‘Yuan, 1983). On the other hand, there is litde evi- dence of the superiority of more complex, con- flict-focused interventions for clients with more complex difficulties. Functional impairment is the observed or rated degree of impairment in daily functioning. ‘The literature is often unclear concerning the cause and effect or even the temporal relationship ‘between symptomatic status and functional impair- ment, although it is often assumed that symptom status leads to various degrees of functional impairment. Even among medical diseases, how- ever, the degree of functional impairment may vary substantially, even in individuals with the same medical pathology or psychiatric condition. Reviews indicate that level of functional impair- ‘ment is negatively correlated with prognosis across disorders such as depression (Gitlin, Swendsen, & Heller, 1995; Kocsis et al., 1988; Sotsky et al., 1991), bulimia nervosa (Fahy & Russel, 1993), obsessive-compulsive disorder (Keijsers, Hoog- duin, & Schaap, 1994), and chemical dependency (McLellan, Woody, Luborsky, O’Brien, & Dru- ley, 1983). If one regards disturbed object rela- tions Goyce & Piper, 1996) and comorbid ~212 * Chapter 6 / The Influence of Client Variables on Psychotherapy personality disorders among substance abusers (Woody et a, 1984) as indications of functional impairment, these have been found to relate neg- atively to psychodynamic treatment outcome. ‘Three client variables—subjective distress, reactance/resistance, and coping styles—describe the way the individual deals with problems and symptoms and thus might be important client variables predicting psychotherapy outcome. Sub- jective distress refers to the client’ internal state rather than objective behavior or performance, and clinically ic is assumed that this internal state would have motivational properties. There is modest support for the assumption that subjective distress is motivational. There is also support for the assumption that psychosocial treatment has its greatest effects on those clients with moderate to high levels of subjective distress (Klerman, Dimascio, Weissman, Prusoff, & Paykel, 1974; Lambert & Bergin, 1983; McLean & Taylor, 1992). In the NIMH Collaborative Study of Depression, those clients with the most severe distress were most effectively treated by IPT, whereas IPT and CBT worked wel for those with mild and moderate distress (Elkin, 1994; Elkin, Gibbons, Shea, & Shaw, 1996; Imber et al., 1990). Since psychotherapy is a situation in which the client can potentially learn from the therapist, the client’ receptivity to information, direction, advice, and interpretation from the therapist may be crucial to treatment success. Reactance is a con- struct defined by describing the behavior of an individual who responds in oppositional ways to perceived loss of choice (Brehm, 1966, 1976; Brehm & Brehm, 1981). Reactance theory is thoughtful discussion of instances in which thoughts and behavior are free and unimpeded as compared to instances of reactance in which an aversive motivational state occurs and autonomous behavior is threatened. Brehm (1976) suggested that reactance might occur in psychotherapy in instances where the client attempts to avoid the influence of the therapist. Psychodynamic ther- apy has often been conceptualized as an effort to understaiid and interpret the resistance of the client. Others in a more cognitive and behav- ioral tradition have suggested that reactance can not only be dealt with, but also utilized in the therapeutic encounter (Tennen & Affleck, 1991; Tennen, Eron, & Rohrbaugh, 1985; Tennen, Rohrbaugh, Press, & White, 1981) to enhance ‘outcome. Reactance and resistance involve a number of client behaviors and attitudes that describe a range of behaviors from simple non- compliance to delayed compliance to oppos.” tional behavior in the face of the therapis authority. It is quite likely that therapeuye impasses as defined by Safran and Muran (2009) often involve instances that could be conceptual. ized as reactance between client and therapi, thus stimulating the investigation of how to man, age and/or utilize these situations. Client resistance has been shown to be aso. ated with poor prognosis with psychothe @ischoff & ‘Tracey, 1995; Miller, Benefield, & Tonigan, 1993; Stoolmiller, Duncan, Bank, & Patterson, 1993). A direct approach to these sty. ations is that of Shoham-Salomon, Avner, and ‘Neeman (1989). Reactance was measured as 4 pretreatment variable by the client's content. filtered tone of voice. In a treatment utilizing paradoxical interventions, those with higher pretreatment reactance benefited more from the therapy than those with lower reactance scores. A self-report measure of reactance (Dowd, Mile, & Wise, 1991) was significantly correlated with traits such as dominance, independence, auton- ‘omy, denial, self-sufficiency, lack of tolerance, snd lack of conformity. In yet another study (Shoham, Bootzin, Rohrbaugh, & Urry, 1996), the role of reactance and treatment for insomnia was exam- ined. Ie was found that paradoxical interventions were more effective for the high-reactance cliens than for the low-reactance clients and that pro- gressive muscle relaxation treatment was mort ‘effective for low-reactance clients. Beutler and colleagues define coping style the conscious and unconscious behaviors that at designed to enhance the individual's ability © avoid the negative effects of anxiety and ro adapt to the environment. There is a body of literature that grossly divides coping styles into those tht are externalizing (e.g., impulsivity, projection, sociopathic behavior) and those that are interma- izing (eg, obsessiveness, inhibition, inner direct edness, and restraint). Clients at varying levels of externalizing and internalizing respond dif ently to various treatments. For example, anf alcoholic subjects, individuals high and low ot externalization/impulsivity responded differe'Y to behavioral and interpersonal treatments. The externalizing clients did better when treated wit behavioral treatments, and the introspective 008 did better with an interpersonal therapy. Simi larly, Longabaugh et al. (1994) found that ale, holies who were externalizing responded better ® cognitive behavioral treatment than they did ® relationship enhancement therapy. These resué, i iBis Nic pis, tan. 0. py L& ak ine a sing sher sh he, and am, of ons a a3 ate * ae were not replicated in the large-scale Project MATCH (1997), which we described elsewhere in this chapter. Among a group of outpatients, cognitive therapy was more effective than inter- sonal therapy among clients who were exter- fulizing, and interpersonal therapy was most effective for the internalizing clients (Barber & Muenz, 1996). ‘Although the clients’ social supports in some ‘rvs external to the cient, it also seems clear that cients play a major role developing (or destroy- in.) social support network composed of friend- ships, work, and other relationships. Social support {sa summary statement about the interpersonal context within which the individual operates and ha: been found to be a potent variable in treat- ment outcome. Social support has been measured 4s both the objective presence of others in the environment and the subjective sense that support is avilable. There is ample evidence that social support, especially the subjective sense of support, provides a buffer against relapse and improves prognosis (George, Blazer, & Hughes, 1989; Hoo- ley & Teasdale, 1989; Longabaugh, Beattie, Noel, Stout, & Malloy, 1993; Moos, 1990; Zlotknick, Shea, Pilkonis, Elkin, & Ryan, 1996). Sets of Client Variables and Generalized Anxiety Disorder Borkovec (Borkovec & Miranda, 1999) has exam- ined client variables in relationship to the suc- cessful treatment of individuals with generalized ansiey disorder (GAD). Given the presence of GAD, these researchers have studied client char- atterstcs relevant tothe disorder itself, including attention, thought, imagery, emotional psycho- Physiology, and their interactions. For example, 3t the physiological level, GAD is characterized ¥ autonomic inflexibility due to a deficiency in Parasympathetic tone. Thought content charac- tered by worry reduces parasympathetic tone. 'n addition, threatening words generate a defen- {Re Mesponse in these clients, serving as an uncon- vonal stimulus that leads to an orientation to sociated conditional stimuli. At the level of tetPersonal behavior, these clients have been fund tobe differen from controls in their attach *nt-related childhood memories. On the Inven- ‘01y of Interpersonal Problems (HP), a self-report Ietsure of areas of interpersonal difficulties, there 3s thee diferent subtypes: (1) overly nurturant intrusive in their interpersonal relations, (2) Rally avoidant and unassertive, and (3) domi- * and hostile. The authors point out that all of Process and Outcome * 213 these client variables are relevant for differential treatment planning with GAD individuals. Aptitude by Treatment Interaction Research Reviews of psychotherapy research _(Kopta, Lueger, Saunders, & Howard, 1999) often con- clude that there is no evidence supporting the attractive notion that the individual client should bbe matched to a particular treatment tailored to that client difficulties and other characteristics. However, the counterargument is that most psychotherapy studies lack sufficient power to examine potential matches between client and psychotherapy. In addition, clinicians work on the assumption that clients should be matched to par- ticular psychotherapies and aspects of psychother- apy and therefore miss important relationships. ‘The research corollary of the clinical attempt to guide treatment selection on the basis of client variables is a design that assesses the interactions of the treatment type or condition with the client ‘variables, so-called aptitude by treatment interac- tion (ATI) research (Cronbach, 1975). Smith and Sechrest (1991) have emphasized the design requirements for a fruitful exploration of appropriately matching clients according to certain aptitudes with specific treatments. They ‘warn that ATIs may be infrequent, undependable, and difficult to detect. The treatment of alcohol and drug addictions has drawn a number of attempts to specify ATIs. Probably one of the most extensive attempts to match client to treatment was done in Project ‘MATCH, involving individuals with alcoholism who were treated with one of three treatments (Connors et al., 2000). For outpatients, ratings of alliance were positively predicted by client age, motivational readiness to change, socialization, and level of perceived social support. Client edu- cational level, level of depression, and meaning seeking were negatively related to alliance. Among. aftercare clients, alliance was positively predicted by readiness to change, socialization, and social and negatively predicted by level of ion. However, of the variables manifesting positive relationships with alliance, only a few were significant predictors in multiple regression equations. For outpatients client age and motiva- tional readiness to change were positive predic- tors, whereas education was a negative predictor of ratings of alliance. Tn reference to matching clients to treat- ments as related to outcome (Project MATCH214 * Chapter 6 / The Influence of Client Variables on Psychotherapy ; Research Group, 19976), 11 client attributes were examined. Alcohol-dependent outpatients, ‘high in anger and treated in motivational enhance- ‘ment therapy, had better post-treatment drink- ing behavior than an analogous group treated with Cognitive Behavioral Coping Skills Ther- apy (CBT). Aftercare clients high in alcohol dependence had better post-treatment outcomes in Twelve-Step Facilitation Therapy, and low- dependent clients did better in CBT: less developed, yet ambitious, project has been reported by Beutler, Moleiro, Malik, and Harwood (2000) to test the effects of a Pre- seribed Therapy against competing therapies for mixed group of clients with substance abuse and depression. The prescriptive treatment focused n tailoring the treatment to four salient client characteristics (described earlier in this chapter): level of functional impairment, internalized or externalized coping, level of reactance, and level of distress. The prescripted treatment matched treatment and therapist characteristics to each of these four client variables: level of functional impairment modified the intensity of treatment, coping was matched to focus on meaning of behavior; reactance was matched with therapist directiveness; and distress was matched to thera- pist support or arousal techniques. A hierarchical analysis suggested that the fit of clientand thera- pist across the three treatment conditions made a modest contribution to predictive power at the end of treatment and a large contribution at the end of a six-month followup period. Much more work is needed, but this research was generated by the plausible yet infrequently researched notion that the therapist should adapt to client variables. This approach is ereative and refresh- ing as compared to the dominant research theme today of matching the client on only the diagno- sis variable to treatments conceptualized in terms of theory and school of psychotherapy. Summary ‘The ATI design has been used infrequently, despite its design benefits. One of the reasons ‘might be that theoretical models may not be suf- ficient to use ATIs, inasmuch as the basic research fon the pathology must be done first. Also, they require the time and expense related to gather information on a large number of clients. ‘The finding of the ATI research to date has been rela- tively disappointing, and Project Match isa prime example. The model of the addiction pathology ‘may have been limited, and therefore the client variables chosen were not central tthe pathy” ogy itself Further research is needed befor abandoning more complete study of clint yt ables and their contribution to outcomes win this paradigm. Conclusions and Implications 41. The field of psychotherapy research hy crystallized around the randomized clinical way for clients “homogeneous” for a particular Ds, IV diagnosis. This research, furthered by NIN and its funding, has been characterized as a Foo} and Drug Administration approach (Pilkonis Krause, 1999), with its goal of establishing hg evidence of treatment safety and efficacy in least two clinical tials. This approach fosten internal validity and provides litte consideratiog of clinical significance. The focus is on trex ‘ments, with ltd attention to patients, therapisy ot individual differences. The yield ofthis orien tation is group outcomes reflected in group meay scores, with no attention to mediators and mod. erators of outcome. This research concentration has led to the “empirically validated treatment movement, which argues that the matching ofthe client variable of diagnosis with a particult treatment should be preferred in clinical pric tice and should be included in the training of clinical psychologists In contrast, we argue that itis precisely this kind of oversimplifcation that leads to the gopin understanding and information exchange between researchers and practitioners. Everyday clinica reality is one in which the diagnosis s only one of many client variables that must be considered in planning a treatment intervention. Nondiag- nostic client characteristics may be more usefil predictors of psychotherapy outcome than DSM- based diagnoses, The diagnostic categories aio for too much heterogeneity in personality tis to serve as useful predictors or matching var ables. Psychotherapy research designs shcald, therefore, stress the interaction between cliet diagnosis and other salient client characters with intervention strategies. : 2. Ifone abandons the simplistic notion tht assessment of client diagnosis alone provides # clear road to treatment, one is faced with an ove whelming number of client variables to considet Ieis impossible to adequately research all the variables in either post-hoc analysis of seat ment studied focused on the brief treatment symptom diagnostic constellation, or in plant (Be inkprospective studies of nondiagnostic client vari- ables. This review is an attempt to bing some order and perspective on the client variables that have shown promise thus far. The field has pro- sgressed from an early focus on client demographic variables to a focus on personality traits/disorders, especially those that are related to the nature of the disorder itself. 3. Single-client variables do not operate alone, as the individual client is a complex integrated person. Thus, research focused on a constellation of salient variables will be likely to show the great- cst impact on treatment process and outcome. ‘The work of Beutler and colleagues, Piper and col- leagues, and Borkovec are exemplary inthis regard. 4. Unfortunately, most of the research on nondiagnostic client variables involves a post-hoc analysis of the impact of various client variables on the outcome of interest. The examination by Sotsky and colleagues of the multisite NIMH col- laborative depression study is an example of this type of investigation It is interesting to contrast this approach with the theory-driven approach of Blatt and colleagues to the same data set. Although both approaches are informative, the field will be likely to make more progress if the latter direction with theory-guided inquiries is used. A further methodological progression is to investigate either individual or sets of nondiag- nostic client variables in a prospective study. The work of Shoham-Salomon and the MATCH studies are prime examples. The most creative approach to date is to articulate areas of client ‘ariability that are likely to have the most power- ful effect on treatment process and outcome, and to match the therapist behavior, regardless of School of psychotherapy, to the needs of the client Beutler, Moleiro, Malik, & Harwood, 2000). Itis in this work that the focus on client variables in interaction with therapist variables rather than looking at isolated variables is brought center Stage and hopefully will result in more progress. 5. Currently, a major research concern is to (tend efficacy research that is conducted on Highly selected clients at research centers with ‘Stefully selected therapists to research that eval- 836 the effectiveness of specific therapeutic wproaches to a more heterogeneous group of ‘lents in the local community treated by com- unity therapists, We agree that the central Peston of the generalizability of results from the Ratewhat pristine circumstances to the more “erogeneous community setting is a crucial Process and Outcome © 215 cone. An essential issue in this transfer has to do with client variables. In an efficacy study, efforts are made to limit and control client and therapist variability. Studies that are aimed at generalizing results will enhance the likelihood of improving outcomes for clients. 6. Most reviews of client variables in rela- tionship to psychotherapy process and outcome are pessimistic because of the inconsistent and less than clear relationships described in the liter- ature (Garfield, 1994; Petry, Tennen, & Affleck, 2000). Such reviews, including this one, must come to terms with this inconsistency in results. ‘There have been a number of plausible problems {n past approaches to client variables: First, as emphatically stated by Smith and Sechrest (1991), a number of design issues must be addressed in order to provide a research set- ting in which client aptitude by treatment inter- actions can be detected, including sufficient numbers of subjects, a clear and theoretically sound articulation of mechanisms of change, and a strong treatment that is of sufficient duration, and intensity to bring about change. Second, pretreatment client variables have a plausible impact on the therapy, but as soon as, therapy begins, the client variables are in a dynamic and ever changing context of therapist variables and behavior. There is a growing aware- ness and articulation of the inherent interactive nature of psychotherapy such that pretreatment client variables will have only a modest and often inconsistent relationship with therapy process and outcome. The therapists responsiveness to client variables and behavior will determine the statistical relationship of the client variable to ‘outcome (Stiles, Honos-Webb, & Surko, 1998). Third, client variables can function in differ- cent ways. Most of the research has attempted t0 isolate single-client variables that have a prognos- tic relationship to therapy process or outcome. Often, reviews are tallies of which studies are pos- itive and which ones are negative on a single- client variable. Often they are post-hoc client variables of convenience rather than theoretically driven explorations. There is often no clear rationale as to whether the client variable is a ‘mediator or moderator. These variables are treated as mediators or moderators based on their char- acteristics, for example, gender or age. Te is important to determine, both theoreti- cally and statistically, whether a particular client variable operates as a mediator, a moderator, or ~440 * Chapter 6/ Lhe intuence of Cuent yaruuics vu rsycnumierayy both (ee Whisman, 1993). Mediating variables are not independent of moderator variables, and vice versa. Whisman also points out that the degree of mediation for a particular variable may be contingent on the level of a given moderator. James and Brett (1984) called this model “moder- ‘ated mediation.” Moderated mediation may be fone reason previous research has often found contradictory results in regard to the relationship of client variables to outcome. “To the extent that a mediating variable i also a moderated variable, it becomes a prescriptive variable. For example, in the NIMH treatment of depression study, severity of illness is not only a prognostic variable but becomes a prescriptive variable because the most severely depressed clients responded to IPT and medication plus clinical management. Furthermore, the work of Borkovec implies that diagnosis is a prescriptive variable but only to the extent that itis moderated by important client variables related and specific to the diagnosis under question ‘The individualized and more general charac- teristies of the clients who come for psychotherapy are central to the clinical enterprise of psychother- apy practice and the research investigation of psy- chotherapy. The focus of psychiotherapy is on the clients’ problem and diagnosis. Diagnosis is a statement based on common elements among many individuals, whereas an individual client’s problem approaches a statement about the specific difficulties that are woven into the fabric of an individval’ life at one point in time. Client charac- teristics are central to motivation for and the nature of participation in psychotherapy. Motiva- tion for change and participation in treatment is individualized in the interaction between a par- ticular therapist and a particular client. Client characteristics that are relevant to interpersonal processes are paramount in understanding the oad to treatment outcome. The progress of psy- chotherapy and its research will depend directly ‘on our efforts to further the exploration, under- standing, and measurement of clients who seek our assistance with the difficulties they face. FOOTNOTES 1. The words client, patient, and consumer are used differently by various professional groups. In this ‘chapter, we use the convention of cen, even though. psychotherapy is now planned, paid for, and re- searched according to a DSM diagnosis inferring patient status. We would point out that all three terms infer a relaonship with another: cine under the protection of or receiving profess advice from an advisor; a patients suffering fromay illness and receives care from 2 doctor; and 2 cog sumer buys services from his or her insurance ple and a managed care provider. 2. Managed care is very interested in identifying ing. viduals who, in their minds, do not need therapy. Thy concept of medical necessity is employed to limite payment for therapy. We have found no research oy the use of the concept of medical necessity REFERENCES Ablon, J. & Jones, EB. (1999). Psychotherapy process in the National Insite of Mental Health Treatment of Depression Collaborative Research Program. Journal of Conculting and Cli ical Prycology, 610), 64-75. Agosti,V, Nunes, E., 8 Ocepeck-Welikson, K. (1996). Client factors related to early attrition from an outlient cocaine research clini. Amer can Journal f Drug and Alebal Ase, 22, 29-9, Ainsworth, M. (1964). Paterns of arachment beha. jor shown by the infant in interaction with is mother. Merrill Palmer Quarterly, 10, 51-58, Alpher,V.S, Perfetto, Henry, W.B, & Strupp, HH, (1990), Dynamic factors ineent assessment and prediction of change in shore-erm dynamic py chotherapy. Pycbutbrapy, 27, 350-361. American Paychiatri Association. (1980). Dingnati ‘and Statiical Manual of Mental Disorders id ed). Washington, DC: American Psychiatric ‘Associaton American Psychiatric Association. (1994). Diagwonic snd Staiical Manual of Mental Dizorder (Ath ed). Washington, DC: American Psychiatric Association, ‘Arean, PA. Perri, M. G., Neou, A. M, Schein, R-Ly Christopher, F, & Joseph, TX. (1993). Compar ative effectiveness of social problem-solving ther apy and reminiscence therapy as treatments for depression in older adults. journal of Connting ‘and Clinical Prycelogy, 61, 1003-1010. Armbruster, P, & Fallon, T (1994). Clinical, sociode smographic, and systems risk factors for atriton ina children’ mental health clinic. Amerian—- Journal of Orbopychisty, 64, 577-585. ‘AuBichon, BG., & Malatesta, V.J. (1994). Obsesie compulsive patients with comorbid personality disorder: associated problems and response t04 comprehensive behavior therapy. Journal of Ci: cal Pybiatry, 55, 448-853. Bacr, P.E., Dunbar, W,, Hamilton, J. E, & Beult! 1. £. (1980). Therapists’ perceptions ofthe psy> shocherpeute proces: Development of therapy process inventory. Pychologica! Reports 4 308310 ne arerensee.FawteJ.P, Consly MB, Crits-Christoph,P, Gladis,L. & Siquelan,'L. (2000). Alzne? pre- dias cients outcome beyond in-trestment change ia symptoms. Journal of Conuling and _Cinieal Pycolgy, 68, 1027-1032. {Barber J.B, & Mena, L. R (1996). The role of avoidance and obsessiveness in matching patients {0 cognitive and interpersonal psychotherapy. Empirical findings from the treatment for depression collaborative research program. Jour nul of Consulting and Cinzl Pchaogs 64, 951-958 Bartow, DH. (1996). The effectiveness of poy- chotherapy Science and policy. Clinical Pyctol- ca: Since and Practice, 3, 36240, Burholomew, K. (1997). Adult tachment processes: Individual and couple perspectives. Bri Journal of Medica! Peboogy, 70, 249-263, “Burholomes, K, & Horowitz, L M. (1991). Atach- ment syles among young adults: A test ofa four- ‘ategory model. Journal of Personality and Seca! 3 Ppebaligy, 61, 236-244, iBecthum, E-E, (1989) Improvement afer evaluation in psychotherapy of depression: Evidence ofa placebo effect? Journal of Chinial Prychelgy, 45, & 945-950, Bergin, AE, & Lambert, MJ. (1979), Counseling 5 he esearch. Councling Poctlogia, 83), 53°56, Berrigan, L. P, & Garfield, SL. (1981). Reladonship of missed psychotherapy appointment to prema: tore termination and social class, Tbe British Jour= tal of Clinical Peles, 20, 239-282. Bees, LE. (1979). Toward specific psychological therapies fo specific conditions. fournl of Can- 5 thing and Clinical Pcbloy, 47, 882-899, Broder, LE, & Berrea, M. R.( 1995). Inegrative ‘iesmen of adle personaly, New Yorks Guilford euler, LE. & Clerkin, J. F190). Sytem tea Mott slecion: Toward targeted terapeatc itercon- Inet! New Yorke Brunnes/Mazel. sewer, LE, Clarkin, JF, & Bongar, B. 2000). 2 Guidlines forthe systematic rate ofthe 2 fovesd ent, New York Oxford Universiy 8 soir Sse Pevls LE. Consol, AJ, & Wilms, RE. (995). Integrative and eclectic therapies in prace pad Bs Bongar & LE. Beutler (Eds), Com Brebensive textbook of pycbotberapy: Theory and ‘Arie (p. 274-282), New York: Oxford Univer- sity Press "is LE, Goodtich, G, Fisher, D., & Williams, 0.8. (1995), Use of psychological tess/nest, (4), The we of pba te for reament. ing and out asesrment nd ed.) (pp. Reet 113) Hillsdale, NJ: Lawrence Exlbause oS: LE. & Hamblin, D1. (1986) Indivdeal- £4 outcome measures of internal change ere i I e = : i References © 217 ‘Methodological considerations Journal of Con- sulting and Clinical Prycbology, 54, 48.53. Beuder,L. E, Moleiro, C, Malik, M, & Harwood, T ‘M. (2000, June). The UC Santa Barbara Study of fircing therapy to patients First resule. A paper presented a the annual meeting ofthe Soy for Peychotherapy Research Chicago,IL. Bovier LF, Sin aan Schrelen,D. Corbishley, A, Hamblin, D, Meredith K, Poren, R, Bamford, C. R,, & Levenson, A. I (1987). (Group cognitive therapy and alprazolam inthe treatment of depression in older adults. Jounal of Consulting and Clinical Prcbology, 55, 550-556, Beutler, L-E,, Wakefcld,P, & Willams, RE (1994). Use ofpoychologicaltest/inseruments for tweatment planing. In M, Marish (El), Ue of coloring for teaten planning rad our, came asesment (pp. §5-14). Chicago: Lawrence Erfbaum, Birkel, RC, & Reppucc, N. D. (1983). Socal net- ‘ora, information-seeking, andthe vlzation of services. American Jeurnal of Community Peta oy, 11, 185-205. Bischoff, M. M, & Tracey, T.J. G. (1995). Client resistance as predicted by therapist behaviors A seudy of sequential dependence: Journal of Coon- ‘cng Pyboog, 42, 487-495. Blatt, S.J, Ford, R.'Q., Berman, W. H., Jr, Cook, B., Cramer, P, & Robins, CE. (194. Tberapeui shange: An objet relations perpecce. New Yorke Plenum Press. Bla, S.J Quinln, D. M, Pilkonis, P..,& Shea, ‘MT. (1995) Impact of perfectionism and need {for approval on the brief treatment of depression ‘The National Institute of Mental Health Treat- ment of Depression Collaborative Research Pro- ram revisited. Journal of Conuting and Cina! Prcboogy, 63, 125-132. Borkovec, T. D.,& Miranda J. (1999). Between sr0up prychotherapy outcome research and basic science. Jura ef Clinical Poly, 55, 147-158, Bowlby, J. (1969), Disruption of affecional bonds and its effects on behavior. Canad’ Mental Heath. Supplement, 58,12. Bowlby J (198), By ethology out of pycho analysis ‘An experiment in interbreeding. Animal Bebos iow, 28, 649-656. Bowlby, J. (1988). A secure base: Parent-child aac ment and belthy buman development. 208), New York: Basie Books, Boyd, J. H. (1986). Use of mental health services for the reaument of panic disorder. American Journal of Pbiaery, 143, 1569-1574, Brehm, J. W. (1966). theory of pocblogical reactance New Yorke Academie Press. Brehm, S. S. (1976), The appt of wal pycoligy to lnzal practice, Washington, DC: Hemisphere Pres. ”218 + Chapter 6 / The Influence of Client Variables on Psychotherapy Brehia, 8. S & Brehm, J. W. (1981), Pocbologialreac- tance: A theory of freedom and contr. New York: ‘Academic Press. Brown, berg, R. G, & Juster, H.R. (1995). Social phobia subtype and avoidane personality, disorder: Effect on severity of socal phobia, impairment, and outcome of cognitive behavioral treatment. Bebavior Therapy, 26, 467-486. Brown, J. A. (1979) Ego identity states and personal- ity variables. Dissertation Absracts Internationa, 40, 1879-1880. Burns, D. D., & Nolen-Hoeksema, S. (1992). Thera- peutic empathy and recovery from depression in cognitive) behavioral therapy: A structural equa~ ‘on model. Jaurnal of Consulting and Clinical Py- ology, 60, 441-449. Butler, T, Giordano, S., & Neren, 8. (1985). Gender and Sex-role attributes as predictors of usization ‘of natural support systems during personal stress evens. Ser Roles, 13, 515-524 Chambless, D. L., Baker, M.J, Buucom, D.H., Beut- ler, L-E., Calhoun, K. S, Crits-Christoph, Daiuto, A, DeRubeis,R., Derweiler, J, Haga, D.A.E,Johnson, S,, MeCury,S., Moeser, K.T, Pope, K.S., Sanderson, W. C., Shoham, V, Stickle, T,, Williams, D. A. & Woody, $. R (1998). Update on empizicaly validated thers- Pies: I. Tee Clinical Prycologi, 51, 3-16. Chambless, D. L, Tran, G. Q., & Glass, C. R (1997). Predictors of response to cognitive-behavioral group therapy for social phobia. Journal of Ansi- 47 Disorder, 11, 221-240. : Chiesa, M., Drahorad, C., & Longo, S. (2000) Early termination of treatment in personality disorder ‘weated in a psychotherapy hospital. British our- anal of Pociatry, 177, 107-111. (Clarkin, J.F, & Abrams, R. (1998). Management of personality disorders in the context of mood and anxiety disorders. In A.J. Rush (Ed), Mood and ‘anxiety diarders (pp, 224-235), Philadelphia: Williams & Wilkins. (Clarkin, J F, Yeomans, F-E., & Kemberg, O. F (1999). Prycbutberspy for borderline peroality. New York: John Wiley & Sons. Clementel Jones, C., Malan, D., & Trauer,‘T: (1990). A retrospective follow-up study of 84 clients ‘teated with individual psychoanalytic psy- chotherapy: Outcome and predictive factors. British Journal of Pychotberay, 6, 363-376. Coker, S. Vize, C., Wade, T, & Cooper, BJ. (1993). Patients with bulimia nervosa who fail to engage in cognitive behavior therapy International Jour- nal of Eating Disorders, 13, 35-0. ‘Connors, G.J., DiClemente, C. C,, Dermen, K. H., Kadden, R,, Carroll, KM, & Frone, M. R (2000), Journal of Seudies on Alcbe, 61, 139-19. Conte, H., Platchik, R, Picard, S., 8 Karas, (1991). (Can personality traits predict psychotherapy out- come? Comprebensce Pchiatry, 32, 66-72. Cooper, PJ, Coker, & Fleming, C. (1994), sy help for bulimia nervoss: A preliminary repo * International Journal of Eating Ditorders, 1 401-404. Contra, Messy, P, Marks, I. M, Molar, Eg Bouvard, (1993). Predictive factors in the ee, ment of obsesive-compulsive disorders with fy vvoramine and/or behaviour therapy. Bebevonny Prycbotherapy, 21, 45-50, Citendon, K S, Manfred C, Lacey, L., Wemede, R, & Parsons, J. (1994). Measuring readiness gy ‘mativason to quit smoking among women in pb Ticheahh linis. ddiaive Behaviors 19, 97-303, Cronbach, LJ (1975). Beyond the two disciplines scientific psychology. American Peychlois, 3, 116-127. Cyranowsbi,J. M, Frank, E., Young, E, & Shes, 'M. K. (2000). Adolescent onset ofthe gender di. ferences in ifetime rates of major depression: A theoretical model. Archive of General bia, 57, 21-27, Diamond, D,, Clarkin, J, Levine, H., Levy, K, Foelsch, B, & Yeomans F. (1999). Atachment theory and borderline personality disorder: A preliminary report. PreboenayticInguiry, 19, 831-884, DiClemente, C. C, & Prochaska, J. 0. (1982). Sele change and therapy change of smoking behaion ‘A comparison of processes of change of cesitinn and maintenance. Additive Bebaviers, 7, 33-182 DiClemente, C. C, Prochaska, J. 0,, Fuishurst, S.K, ‘Velicer, W. F, Velasques, M. M,, & Rossi JS, (4991), The process of smoking cessation: An analysis of precontemplation, contemplation, al preparation stages of change. Journal of Conuing and Clinical Prychology, 58, 295-308. Diguer, L., Barber, J.P, & Luborsky, L. (1993). Thre concomitants: personality disorders, psychiatric severity and outcome of dynamic paychotherspy (of major depression. American Journal of Pychi- 19, 150, 1246-1248. Dobson, KS. (1989). A meta-analysis ofthe effiney of cognitive therapy for depression. Journal of Consulting and Clinical Phycology, 57, 414-419. Doda, J. (1970). A retrospective analysis of vati- ables related to duration of treatment in a uriver sity elni. Journal of Nervou and Mental Diz, 151, 75-85. Dolan, R. T. Arnkoff, D.B. & Glass, C. R. (1993) Client attachment style and the psychotherapists interpersonal stance. Pachatherapy, 30, 408-412. Dowd, E.T, Milne, C.R, & Wise, 8. L. (1991). Tae ‘Therapeutic Reactance Seale: A measure ofp} chological reactance. Journal of Coumelng and Development, 68, 541-545. Dozier, M. (1990). Atachment organization and ws ‘ment use for adults with seriou psychopatholog- cal disorders. Development and Prychepatblogy, 2 47-60. w bs rh thi i i nd Eye Fahy Fat